intl/icu/source/i18n/quant.cpp@6474c204b198 (annotated)
intl/icu/source/i18n/quant.cpp
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 | ********************************************************************** |
| michael@0 | 3 | * Copyright (C) 2001-2012, International Business Machines |
| michael@0 | 4 | * Corporation and others. All Rights Reserved. |
| michael@0 | 5 | ********************************************************************** |
| michael@0 | 6 | * Date Name Description |
| michael@0 | 7 | * 07/26/01 aliu Creation. |
| michael@0 | 8 | ********************************************************************** |
| michael@0 | 9 | */ |
| michael@0 | 10 | |
| michael@0 | 11 | #include "unicode/utypes.h" |
| michael@0 | 12 | |
| michael@0 | 13 | #if !UCONFIG_NO_TRANSLITERATION |
| michael@0 | 14 | |
| michael@0 | 15 | #include "quant.h" |
| michael@0 | 16 | #include "unicode/unistr.h" |
| michael@0 | 17 | #include "util.h" |
| michael@0 | 18 | |
| michael@0 | 19 | U_NAMESPACE_BEGIN |
| michael@0 | 20 | |
| michael@0 | 21 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(Quantifier) |
| michael@0 | 22 | |
| michael@0 | 23 | Quantifier::Quantifier(UnicodeFunctor *adoptedMatcher, |
| michael@0 | 24 | uint32_t _minCount, uint32_t _maxCount) { |
| michael@0 | 25 | // assert(adopted != 0); |
| michael@0 | 26 | // assert(minCount <= maxCount); |
| michael@0 | 27 | matcher = adoptedMatcher; |
| michael@0 | 28 | this->minCount = _minCount; |
| michael@0 | 29 | this->maxCount = _maxCount; |
| michael@0 | 30 | } |
| michael@0 | 31 | |
| michael@0 | 32 | Quantifier::Quantifier(const Quantifier& o) : |
| michael@0 | 33 | UnicodeFunctor(o), |
| michael@0 | 34 | UnicodeMatcher(o), |
| michael@0 | 35 | matcher(o.matcher->clone()), |
| michael@0 | 36 | minCount(o.minCount), |
| michael@0 | 37 | maxCount(o.maxCount) |
| michael@0 | 38 | { |
| michael@0 | 39 | } |
| michael@0 | 40 | |
| michael@0 | 41 | Quantifier::~Quantifier() { |
| michael@0 | 42 | delete matcher; |
| michael@0 | 43 | } |
| michael@0 | 44 | |
| michael@0 | 45 | /** |
| michael@0 | 46 | * Implement UnicodeFunctor |
| michael@0 | 47 | */ |
| michael@0 | 48 | UnicodeFunctor* Quantifier::clone() const { |
| michael@0 | 49 | return new Quantifier(*this); |
| michael@0 | 50 | } |
| michael@0 | 51 | |
| michael@0 | 52 | /** |
| michael@0 | 53 | * UnicodeFunctor API. Cast 'this' to a UnicodeMatcher* pointer |
| michael@0 | 54 | * and return the pointer. |
| michael@0 | 55 | */ |
| michael@0 | 56 | UnicodeMatcher* Quantifier::toMatcher() const { |
| michael@0 | 57 | Quantifier *nonconst_this = const_cast<Quantifier *>(this); |
| michael@0 | 58 | UnicodeMatcher *nonconst_base = static_cast<UnicodeMatcher *>(nonconst_this); |
| michael@0 | 59 | |
| michael@0 | 60 | return nonconst_base; |
| michael@0 | 61 | } |
| michael@0 | 62 | |
| michael@0 | 63 | UMatchDegree Quantifier::matches(const Replaceable& text, |
| michael@0 | 64 | int32_t& offset, |
| michael@0 | 65 | int32_t limit, |
| michael@0 | 66 | UBool incremental) { |
| michael@0 | 67 | int32_t start = offset; |
| michael@0 | 68 | uint32_t count = 0; |
| michael@0 | 69 | while (count < maxCount) { |
| michael@0 | 70 | int32_t pos = offset; |
| michael@0 | 71 | UMatchDegree m = matcher->toMatcher()->matches(text, offset, limit, incremental); |
| michael@0 | 72 | if (m == U_MATCH) { |
| michael@0 | 73 | ++count; |
| michael@0 | 74 | if (pos == offset) { |
| michael@0 | 75 | // If offset has not moved we have a zero-width match. |
| michael@0 | 76 | // Don't keep matching it infinitely. |
| michael@0 | 77 | break; |
| michael@0 | 78 | } |
| michael@0 | 79 | } else if (incremental && m == U_PARTIAL_MATCH) { |
| michael@0 | 80 | return U_PARTIAL_MATCH; |
| michael@0 | 81 | } else { |
| michael@0 | 82 | break; |
| michael@0 | 83 | } |
| michael@0 | 84 | } |
| michael@0 | 85 | if (incremental && offset == limit) { |
| michael@0 | 86 | return U_PARTIAL_MATCH; |
| michael@0 | 87 | } |
| michael@0 | 88 | if (count >= minCount) { |
| michael@0 | 89 | return U_MATCH; |
| michael@0 | 90 | } |
| michael@0 | 91 | offset = start; |
| michael@0 | 92 | return U_MISMATCH; |
| michael@0 | 93 | } |
| michael@0 | 94 | |
| michael@0 | 95 | /** |
| michael@0 | 96 | * Implement UnicodeMatcher |
| michael@0 | 97 | */ |
| michael@0 | 98 | UnicodeString& Quantifier::toPattern(UnicodeString& result, |
| michael@0 | 99 | UBool escapeUnprintable) const { |
| michael@0 | 100 | result.truncate(0); |
| michael@0 | 101 | matcher->toMatcher()->toPattern(result, escapeUnprintable); |
| michael@0 | 102 | if (minCount == 0) { |
| michael@0 | 103 | if (maxCount == 1) { |
| michael@0 | 104 | return result.append((UChar)63); /*?*/ |
| michael@0 | 105 | } else if (maxCount == MAX) { |
| michael@0 | 106 | return result.append((UChar)42); /***/ |
| michael@0 | 107 | } |
| michael@0 | 108 | // else fall through |
| michael@0 | 109 | } else if (minCount == 1 && maxCount == MAX) { |
| michael@0 | 110 | return result.append((UChar)43); /*+*/ |
| michael@0 | 111 | } |
| michael@0 | 112 | result.append((UChar)123); /*{*/ |
| michael@0 | 113 | ICU_Utility::appendNumber(result, minCount); |
| michael@0 | 114 | result.append((UChar)44); /*,*/ |
| michael@0 | 115 | if (maxCount != MAX) { |
| michael@0 | 116 | ICU_Utility::appendNumber(result, maxCount); |
| michael@0 | 117 | } |
| michael@0 | 118 | result.append((UChar)125); /*}*/ |
| michael@0 | 119 | return result; |
| michael@0 | 120 | } |
| michael@0 | 121 | |
| michael@0 | 122 | /** |
| michael@0 | 123 | * Implement UnicodeMatcher |
| michael@0 | 124 | */ |
| michael@0 | 125 | UBool Quantifier::matchesIndexValue(uint8_t v) const { |
| michael@0 | 126 | return (minCount == 0) || matcher->toMatcher()->matchesIndexValue(v); |
| michael@0 | 127 | } |
| michael@0 | 128 | |
| michael@0 | 129 | /** |
| michael@0 | 130 | * Implement UnicodeMatcher |
| michael@0 | 131 | */ |
| michael@0 | 132 | void Quantifier::addMatchSetTo(UnicodeSet& toUnionTo) const { |
| michael@0 | 133 | if (maxCount > 0) { |
| michael@0 | 134 | matcher->toMatcher()->addMatchSetTo(toUnionTo); |
| michael@0 | 135 | } |
| michael@0 | 136 | } |
| michael@0 | 137 | |
| michael@0 | 138 | /** |
| michael@0 | 139 | * Implement UnicodeFunctor |
| michael@0 | 140 | */ |
| michael@0 | 141 | void Quantifier::setData(const TransliterationRuleData* d) { |
| michael@0 | 142 | matcher->setData(d); |
| michael@0 | 143 | } |
| michael@0 | 144 | |
| michael@0 | 145 | U_NAMESPACE_END |
| michael@0 | 146 | |
| michael@0 | 147 | #endif /* #if !UCONFIG_NO_TRANSLITERATION */ |
| michael@0 | 148 | |
| michael@0 | 149 | //eof |
