4:8
  +   #ˆl–äY>””†»€jà€¦-ÿÃ„h¶÷_†I|¥+ÁÇ ?÷     #// Â© 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
*
*   Copyright (C) 2002-2014, International Business Machines
*   Corporation and others.  All Rights Reserved.
*
*******************************************************************************
*   file name:  uset.h
*   encoding:   UTF-8
*   tab size:   8 (not used)
*   indentation:4
*
*   created on: 2002mar07
*   created by: Markus W. Scherer
*
*   C version of UnicodeSet.
*/


/**
 * \file
 * \brief C API: Unicode Set
 *
 * <p>This is a C wrapper around the C++ UnicodeSet class.</p>
 */

#ifndef __USET_H__
#define __USET_H__

#include "unicode/utypes.h"
#include "unicode/uchar.h"
#include "unicode/localpointer.h"

#ifndef USET_DEFINED

#ifndef U_IN_DOXYGEN
#define USET_DEFINED
#endif
/**
 * USet is the C API type corresponding to C++ class UnicodeSet.
 * Use the uset_* API to manipulate.  Create with
 * uset_open*, and destroy with uset_close.
 * @stable ICU 2.4
 */
typedef struct USet USet;
#endif

/**
 * Bitmask values to be passed to uset_openPatternOptions() or
 * uset_applyPattern() taking an option parameter.
 * @stable ICU 2.4
 */
enum {
    /**
     * Ignore white space within patterns unless quoted or escaped.
     * @stable ICU 2.4
     */
    USET_IGNORE_SPACE = 1,  

    /**
     * Enable case insensitive matching.  E.g., "[ab]" with this flag
     * will match 'a', 'A', 'b', and 'B'.  "[^ab]" with this flag will
     * match all except 'a', 'A', 'b', and 'B'. This performs a full
     * closure over case mappings, e.g. U+017F for s.
     *
     * The resulting set is a superset of the input for the code points but
     * not for the strings.
     * It performs a case mapping closure of the code points and adds
     * full case folding strings for the code points, and reduces strings of
     * the original set to their full case folding equivalents.
     *
     * This is designed for case-insensitive matches, for example
     * in regular expressions. The full code point case closure allows checking of
     * an input character directly against the closure set.
     * Strings are matched by comparing the case-folded form from the closure
     * set with an incremental case folding of the string in question.
     *
     * The closure set will also contain single code points if the original
     * set contained case-equivalent strings (like U+00DF for "ss" or "Ss" etc.).
     * This is not necessary (that is, redundant) for the above matching method
     * but results in the same closure sets regardless of whether the original
     * set contained the code point or a string.
     *
     * @stable ICU 2.4
     */
    USET_CASE_INSENSITIVE = 2,  

    /**
     * Enable case insensitive matching.  E.g., "[ab]" with this flag
     * will match 'a', 'A', 'b', and 'B'.  "[^ab]" with this flag will
     * match all except 'a', 'A', 'b', and 'B'. This adds the lower-,
     * title-, and uppercase mappings as well as the case folding
     * of each existing element in the set.
     * @stable ICU 3.2
     */
    USET_ADD_CASE_MAPPINGS = 4
};

/**
 * Argument values for whether span() and similar functions continue while
 * the current character is contained vs. not contained in the set.
 *
 * The functionality is straightforward for sets with only single code points,
 * without strings (which is the common case):
 * - USET_SPAN_CONTAINED and USET_SPAN_SIMPLE work the same.
 * - USET_SPAN_CONTAINED and USET_SPAN_SIMPLE are inverses of USET_SPAN_NOT_CONTAINED.
 * - span() and spanBack() partition any string the same way when
 *   alternating between span(USET_SPAN_NOT_CONTAINED) and
 *   span(either "contained" condition).
 * - Using a complemented (inverted) set and the opposite span conditions
 *   yields the same results.
 *
 * When a set contains multi-code point strings, then these statements may not
 * be true, depending on the strings in the set (for exam