intl/icu/source/common/unicode/unimatch.h@6474c204b198 (annotated)
intl/icu/source/common/unicode/unimatch.h
Wed, 31 Dec 2014 06:09:35 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 06:09:35 +0100
- changeset 0
- 6474c204b198
- permissions
- -rw-r--r--
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 | * Copyright (C) 2001-2005, International Business Machines Corporation and others. All Rights Reserved. |
| michael@0 | 3 | ********************************************************************** |
| michael@0 | 4 | * Date Name Description |
| michael@0 | 5 | * 07/18/01 aliu Creation. |
| michael@0 | 6 | ********************************************************************** |
| michael@0 | 7 | */ |
| michael@0 | 8 | #ifndef UNIMATCH_H |
| michael@0 | 9 | #define UNIMATCH_H |
| michael@0 | 10 | |
| michael@0 | 11 | #include "unicode/utypes.h" |
| michael@0 | 12 | |
| michael@0 | 13 | /** |
| michael@0 | 14 | * \file |
| michael@0 | 15 | * \brief C++ API: Unicode Matcher |
| michael@0 | 16 | */ |
| michael@0 | 17 | |
| michael@0 | 18 | |
| michael@0 | 19 | U_NAMESPACE_BEGIN |
| michael@0 | 20 | |
| michael@0 | 21 | class Replaceable; |
| michael@0 | 22 | class UnicodeString; |
| michael@0 | 23 | class UnicodeSet; |
| michael@0 | 24 | |
| michael@0 | 25 | /** |
| michael@0 | 26 | * Constants returned by <code>UnicodeMatcher::matches()</code> |
| michael@0 | 27 | * indicating the degree of match. |
| michael@0 | 28 | * @stable ICU 2.4 |
| michael@0 | 29 | */ |
| michael@0 | 30 | enum UMatchDegree { |
| michael@0 | 31 | /** |
| michael@0 | 32 | * Constant returned by <code>matches()</code> indicating a |
| michael@0 | 33 | * mismatch between the text and this matcher. The text contains |
| michael@0 | 34 | * a character which does not match, or the text does not contain |
| michael@0 | 35 | * all desired characters for a non-incremental match. |
| michael@0 | 36 | * @stable ICU 2.4 |
| michael@0 | 37 | */ |
| michael@0 | 38 | U_MISMATCH, |
| michael@0 | 39 | |
| michael@0 | 40 | /** |
| michael@0 | 41 | * Constant returned by <code>matches()</code> indicating a |
| michael@0 | 42 | * partial match between the text and this matcher. This value is |
| michael@0 | 43 | * only returned for incremental match operations. All characters |
| michael@0 | 44 | * of the text match, but more characters are required for a |
| michael@0 | 45 | * complete match. Alternatively, for variable-length matchers, |
| michael@0 | 46 | * all characters of the text match, and if more characters were |
| michael@0 | 47 | * supplied at limit, they might also match. |
| michael@0 | 48 | * @stable ICU 2.4 |
| michael@0 | 49 | */ |
| michael@0 | 50 | U_PARTIAL_MATCH, |
| michael@0 | 51 | |
| michael@0 | 52 | /** |
| michael@0 | 53 | * Constant returned by <code>matches()</code> indicating a |
| michael@0 | 54 | * complete match between the text and this matcher. For an |
| michael@0 | 55 | * incremental variable-length match, this value is returned if |
| michael@0 | 56 | * the given text matches, and it is known that additional |
| michael@0 | 57 | * characters would not alter the extent of the match. |
| michael@0 | 58 | * @stable ICU 2.4 |
| michael@0 | 59 | */ |
| michael@0 | 60 | U_MATCH |
| michael@0 | 61 | }; |
| michael@0 | 62 | |
| michael@0 | 63 | /** |
| michael@0 | 64 | * <code>UnicodeMatcher</code> defines a protocol for objects that can |
| michael@0 | 65 | * match a range of characters in a Replaceable string. |
| michael@0 | 66 | * @stable ICU 2.4 |
| michael@0 | 67 | */ |
| michael@0 | 68 | class U_COMMON_API UnicodeMatcher /* not : public UObject because this is an interface/mixin class */ { |
| michael@0 | 69 | |
| michael@0 | 70 | public: |
| michael@0 | 71 | /** |
| michael@0 | 72 | * Destructor. |
| michael@0 | 73 | * @stable ICU 2.4 |
| michael@0 | 74 | */ |
| michael@0 | 75 | virtual ~UnicodeMatcher(); |
| michael@0 | 76 | |
| michael@0 | 77 | /** |
| michael@0 | 78 | * Return a UMatchDegree value indicating the degree of match for |
| michael@0 | 79 | * the given text at the given offset. Zero, one, or more |
| michael@0 | 80 | * characters may be matched. |
| michael@0 | 81 | * |
| michael@0 | 82 | * Matching in the forward direction is indicated by limit > |
| michael@0 | 83 | * offset. Characters from offset forwards to limit-1 will be |
| michael@0 | 84 | * considered for matching. |
| michael@0 | 85 | * |
| michael@0 | 86 | * Matching in the reverse direction is indicated by limit < |
| michael@0 | 87 | * offset. Characters from offset backwards to limit+1 will be |
| michael@0 | 88 | * considered for matching. |
| michael@0 | 89 | * |
| michael@0 | 90 | * If limit == offset then the only match possible is a zero |
| michael@0 | 91 | * character match (which subclasses may implement if desired). |
| michael@0 | 92 | * |
| michael@0 | 93 | * As a side effect, advance the offset parameter to the limit of |
| michael@0 | 94 | * the matched substring. In the forward direction, this will be |
| michael@0 | 95 | * the index of the last matched character plus one. In the |
| michael@0 | 96 | * reverse direction, this will be the index of the last matched |
| michael@0 | 97 | * character minus one. |
| michael@0 | 98 | * |
| michael@0 | 99 | * <p>Note: This method is not const because some classes may |
| michael@0 | 100 | * modify their state as the result of a match. |
| michael@0 | 101 | * |
| michael@0 | 102 | * @param text the text to be matched |
| michael@0 | 103 | * @param offset on input, the index into text at which to begin |
| michael@0 | 104 | * matching. On output, the limit of the matched text. The |
| michael@0 | 105 | * number of matched characters is the output value of offset |
| michael@0 | 106 | * minus the input value. Offset should always point to the |
| michael@0 | 107 | * HIGH SURROGATE (leading code unit) of a pair of surrogates, |
| michael@0 | 108 | * both on entry and upon return. |
| michael@0 | 109 | * @param limit the limit index of text to be matched. Greater |
| michael@0 | 110 | * than offset for a forward direction match, less than offset for |
| michael@0 | 111 | * a backward direction match. The last character to be |
| michael@0 | 112 | * considered for matching will be text.charAt(limit-1) in the |
| michael@0 | 113 | * forward direction or text.charAt(limit+1) in the backward |
| michael@0 | 114 | * direction. |
| michael@0 | 115 | * @param incremental if TRUE, then assume further characters may |
| michael@0 | 116 | * be inserted at limit and check for partial matching. Otherwise |
| michael@0 | 117 | * assume the text as given is complete. |
| michael@0 | 118 | * @return a match degree value indicating a full match, a partial |
| michael@0 | 119 | * match, or a mismatch. If incremental is FALSE then |
| michael@0 | 120 | * U_PARTIAL_MATCH should never be returned. |
| michael@0 | 121 | * @stable ICU 2.4 |
| michael@0 | 122 | */ |
| michael@0 | 123 | virtual UMatchDegree matches(const Replaceable& text, |
| michael@0 | 124 | int32_t& offset, |
| michael@0 | 125 | int32_t limit, |
| michael@0 | 126 | UBool incremental) = 0; |
| michael@0 | 127 | |
| michael@0 | 128 | /** |
| michael@0 | 129 | * Returns a string representation of this matcher. If the result of |
| michael@0 | 130 | * calling this function is passed to the appropriate parser, it |
| michael@0 | 131 | * will produce another matcher that is equal to this one. |
| michael@0 | 132 | * @param result the string to receive the pattern. Previous |
| michael@0 | 133 | * contents will be deleted. |
| michael@0 | 134 | * @param escapeUnprintable if TRUE then convert unprintable |
| michael@0 | 135 | * character to their hex escape representations, \\uxxxx or |
| michael@0 | 136 | * \\Uxxxxxxxx. Unprintable characters are those other than |
| michael@0 | 137 | * U+000A, U+0020..U+007E. |
| michael@0 | 138 | * @stable ICU 2.4 |
| michael@0 | 139 | */ |
| michael@0 | 140 | virtual UnicodeString& toPattern(UnicodeString& result, |
| michael@0 | 141 | UBool escapeUnprintable = FALSE) const = 0; |
| michael@0 | 142 | |
| michael@0 | 143 | /** |
| michael@0 | 144 | * Returns TRUE if this matcher will match a character c, where c |
| michael@0 | 145 | * & 0xFF == v, at offset, in the forward direction (with limit > |
| michael@0 | 146 | * offset). This is used by <tt>RuleBasedTransliterator</tt> for |
| michael@0 | 147 | * indexing. |
| michael@0 | 148 | * @stable ICU 2.4 |
| michael@0 | 149 | */ |
| michael@0 | 150 | virtual UBool matchesIndexValue(uint8_t v) const = 0; |
| michael@0 | 151 | |
| michael@0 | 152 | /** |
| michael@0 | 153 | * Union the set of all characters that may be matched by this object |
| michael@0 | 154 | * into the given set. |
| michael@0 | 155 | * @param toUnionTo the set into which to union the source characters |
| michael@0 | 156 | * @stable ICU 2.4 |
| michael@0 | 157 | */ |
| michael@0 | 158 | virtual void addMatchSetTo(UnicodeSet& toUnionTo) const = 0; |
| michael@0 | 159 | }; |
| michael@0 | 160 | |
| michael@0 | 161 | U_NAMESPACE_END |
| michael@0 | 162 | |
| michael@0 | 163 | #endif |
