| 1 | // © 2016 and later: Unicode, Inc. and others. |
| 2 | // License & terms of use: http://www.unicode.org/copyright.html |
| 3 | /* |
| 4 | ********************************************************************** |
| 5 | * Copyright (C) 2001-2011, International Business Machines |
| 6 | * Corporation and others. All Rights Reserved. |
| 7 | ********************************************************************** |
| 8 | * Date Name Description |
| 9 | * 05/24/01 aliu Creation. |
| 10 | ********************************************************************** |
| 11 | */ |
| 12 | |
| 13 | #include "unicode/utypes.h" |
| 14 | |
| 15 | #if !UCONFIG_NO_TRANSLITERATION |
| 16 | |
| 17 | #include "unicode/uchar.h" |
| 18 | #include "unicode/uniset.h" |
| 19 | #include "unicode/ustring.h" |
| 20 | #include "unicode/utf16.h" |
| 21 | #include "titletrn.h" |
| 22 | #include "umutex.h" |
| 23 | #include "ucase.h" |
| 24 | #include "cpputils.h" |
| 25 | |
| 26 | U_NAMESPACE_BEGIN |
| 27 | |
| 28 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TitlecaseTransliterator) |
| 29 | |
| 30 | TitlecaseTransliterator::TitlecaseTransliterator() : |
| 31 | CaseMapTransliterator(UNICODE_STRING("Any-Title" , 9), NULL) |
| 32 | { |
| 33 | // Need to look back 2 characters in the case of "can't" |
| 34 | setMaximumContextLength(2); |
| 35 | } |
| 36 | |
| 37 | /** |
| 38 | * Destructor. |
| 39 | */ |
| 40 | TitlecaseTransliterator::~TitlecaseTransliterator() { |
| 41 | } |
| 42 | |
| 43 | /** |
| 44 | * Copy constructor. |
| 45 | */ |
| 46 | TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator& o) : |
| 47 | CaseMapTransliterator(o) |
| 48 | { |
| 49 | } |
| 50 | |
| 51 | /** |
| 52 | * Assignment operator. |
| 53 | */ |
| 54 | /*TitlecaseTransliterator& TitlecaseTransliterator::operator=( |
| 55 | const TitlecaseTransliterator& o) { |
| 56 | CaseMapTransliterator::operator=(o); |
| 57 | return *this; |
| 58 | }*/ |
| 59 | |
| 60 | /** |
| 61 | * Transliterator API. |
| 62 | */ |
| 63 | TitlecaseTransliterator* TitlecaseTransliterator::clone() const { |
| 64 | return new TitlecaseTransliterator(*this); |
| 65 | } |
| 66 | |
| 67 | /** |
| 68 | * Implements {@link Transliterator#handleTransliterate}. |
| 69 | */ |
| 70 | void TitlecaseTransliterator::handleTransliterate( |
| 71 | Replaceable& text, UTransPosition& offsets, |
| 72 | UBool isIncremental) const |
| 73 | { |
| 74 | // TODO reimplement, see ustrcase.c |
| 75 | // using a real word break iterator |
| 76 | // instead of just looking for a transition between cased and uncased characters |
| 77 | // call CaseMapTransliterator::handleTransliterate() for lowercasing? (set fMap) |
| 78 | // needs to take isIncremental into account because case mappings are context-sensitive |
| 79 | // also detect when lowercasing function did not finish because of context |
| 80 | |
| 81 | if (offsets.start >= offsets.limit) { |
| 82 | return; |
| 83 | } |
| 84 | |
| 85 | // case type: >0 cased (UCASE_LOWER etc.) ==0 uncased <0 case-ignorable |
| 86 | int32_t type; |
| 87 | |
| 88 | // Our mode; we are either converting letter toTitle or |
| 89 | // toLower. |
| 90 | UBool doTitle = TRUE; |
| 91 | |
| 92 | // Determine if there is a preceding context of cased case-ignorable*, |
| 93 | // in which case we want to start in toLower mode. If the |
| 94 | // prior context is anything else (including empty) then start |
| 95 | // in toTitle mode. |
| 96 | UChar32 c; |
| 97 | int32_t start; |
| 98 | for (start = offsets.start - 1; start >= offsets.contextStart; start -= U16_LENGTH(c)) { |
| 99 | c = text.char32At(start); |
| 100 | type=ucase_getTypeOrIgnorable(c); |
| 101 | if(type>0) { // cased |
| 102 | doTitle=FALSE; |
| 103 | break; |
| 104 | } else if(type==0) { // uncased but not ignorable |
| 105 | break; |
| 106 | } |
| 107 | // else (type<0) case-ignorable: continue |
| 108 | } |
| 109 | |
| 110 | // Convert things after a cased character toLower; things |
| 111 | // after an uncased, non-case-ignorable character toTitle. Case-ignorable |
| 112 | // characters are copied directly and do not change the mode. |
| 113 | UCaseContext csc; |
| 114 | uprv_memset(&csc, 0, sizeof(csc)); |
| 115 | csc.p = &text; |
| 116 | csc.start = offsets.contextStart; |
| 117 | csc.limit = offsets.contextLimit; |
| 118 | |
| 119 | UnicodeString tmp; |
| 120 | const UChar *s; |
| 121 | int32_t textPos, delta, result; |
| 122 | |
| 123 | for(textPos=offsets.start; textPos<offsets.limit;) { |
| 124 | csc.cpStart=textPos; |
| 125 | c=text.char32At(textPos); |
| 126 | csc.cpLimit=textPos+=U16_LENGTH(c); |
| 127 | |
| 128 | type=ucase_getTypeOrIgnorable(c); |
| 129 | if(type>=0) { // not case-ignorable |
| 130 | if(doTitle) { |
| 131 | result=ucase_toFullTitle(c, utrans_rep_caseContextIterator, &csc, &s, UCASE_LOC_ROOT); |
| 132 | } else { |
| 133 | result=ucase_toFullLower(c, utrans_rep_caseContextIterator, &csc, &s, UCASE_LOC_ROOT); |
| 134 | } |
| 135 | doTitle = (UBool)(type==0); // doTitle=isUncased |
| 136 | |
| 137 | if(csc.b1 && isIncremental) { |
| 138 | // fMap() tried to look beyond the context limit |
| 139 | // wait for more input |
| 140 | offsets.start=csc.cpStart; |
| 141 | return; |
| 142 | } |
| 143 | |
| 144 | if(result>=0) { |
| 145 | // replace the current code point with its full case mapping result |
| 146 | // see UCASE_MAX_STRING_LENGTH |
| 147 | if(result<=UCASE_MAX_STRING_LENGTH) { |
| 148 | // string s[result] |
| 149 | tmp.setTo(FALSE, s, result); |
| 150 | delta=result-U16_LENGTH(c); |
| 151 | } else { |
| 152 | // single code point |
| 153 | tmp.setTo(result); |
| 154 | delta=tmp.length()-U16_LENGTH(c); |
| 155 | } |
| 156 | text.handleReplaceBetween(csc.cpStart, textPos, tmp); |
| 157 | if(delta!=0) { |
| 158 | textPos+=delta; |
| 159 | csc.limit=offsets.contextLimit+=delta; |
| 160 | offsets.limit+=delta; |
| 161 | } |
| 162 | } |
| 163 | } |
| 164 | } |
| 165 | offsets.start=textPos; |
| 166 | } |
| 167 | |
| 168 | U_NAMESPACE_END |
| 169 | |
| 170 | #endif /* #if !UCONFIG_NO_TRANSLITERATION */ |
| 171 | |