2 **********************************************************************
3 * Copyright (C) 2001-2011, International Business Machines
4 * Corporation and others. All Rights Reserved.
5 **********************************************************************
6 * Date Name Description
7 * 05/24/01 aliu Creation.
8 **********************************************************************
11 #include "unicode/utypes.h"
13 #if !UCONFIG_NO_TRANSLITERATION
15 #include "unicode/uchar.h"
16 #include "unicode/uniset.h"
17 #include "unicode/ustring.h"
18 #include "unicode/utf16.h"
26 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TitlecaseTransliterator
)
28 TitlecaseTransliterator::TitlecaseTransliterator() :
29 CaseMapTransliterator(UNICODE_STRING("Any-Title", 9), NULL
)
31 // Need to look back 2 characters in the case of "can't"
32 setMaximumContextLength(2);
38 TitlecaseTransliterator::~TitlecaseTransliterator() {
44 TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator
& o
) :
45 CaseMapTransliterator(o
)
50 * Assignment operator.
52 /*TitlecaseTransliterator& TitlecaseTransliterator::operator=(
53 const TitlecaseTransliterator& o) {
54 CaseMapTransliterator::operator=(o);
61 Transliterator
* TitlecaseTransliterator::clone(void) const {
62 return new TitlecaseTransliterator(*this);
66 * Implements {@link Transliterator#handleTransliterate}.
68 void TitlecaseTransliterator::handleTransliterate(
69 Replaceable
& text
, UTransPosition
& offsets
,
70 UBool isIncremental
) const
72 // TODO reimplement, see ustrcase.c
73 // using a real word break iterator
74 // instead of just looking for a transition between cased and uncased characters
75 // call CaseMapTransliterator::handleTransliterate() for lowercasing? (set fMap)
76 // needs to take isIncremental into account because case mappings are context-sensitive
77 // also detect when lowercasing function did not finish because of context
79 if (offsets
.start
>= offsets
.limit
) {
83 // case type: >0 cased (UCASE_LOWER etc.) ==0 uncased <0 case-ignorable
86 // Our mode; we are either converting letter toTitle or
90 // Determine if there is a preceding context of cased case-ignorable*,
91 // in which case we want to start in toLower mode. If the
92 // prior context is anything else (including empty) then start
96 for (start
= offsets
.start
- 1; start
>= offsets
.contextStart
; start
-= U16_LENGTH(c
)) {
97 c
= text
.char32At(start
);
98 type
=ucase_getTypeOrIgnorable(fCsp
, c
);
102 } else if(type
==0) { // uncased but not ignorable
105 // else (type<0) case-ignorable: continue
108 // Convert things after a cased character toLower; things
109 // after an uncased, non-case-ignorable character toTitle. Case-ignorable
110 // characters are copied directly and do not change the mode.
112 uprv_memset(&csc
, 0, sizeof(csc
));
114 csc
.start
= offsets
.contextStart
;
115 csc
.limit
= offsets
.contextLimit
;
119 int32_t textPos
, delta
, result
, locCache
=0;
121 for(textPos
=offsets
.start
; textPos
<offsets
.limit
;) {
123 c
=text
.char32At(textPos
);
124 csc
.cpLimit
=textPos
+=U16_LENGTH(c
);
126 type
=ucase_getTypeOrIgnorable(fCsp
, c
);
127 if(type
>=0) { // not case-ignorable
129 result
=ucase_toFullTitle(fCsp
, c
, utrans_rep_caseContextIterator
, &csc
, &s
, "", &locCache
);
131 result
=ucase_toFullLower(fCsp
, c
, utrans_rep_caseContextIterator
, &csc
, &s
, "", &locCache
);
133 doTitle
= (UBool
)(type
==0); // doTitle=isUncased
135 if(csc
.b1
&& isIncremental
) {
136 // fMap() tried to look beyond the context limit
137 // wait for more input
138 offsets
.start
=csc
.cpStart
;
143 // replace the current code point with its full case mapping result
144 // see UCASE_MAX_STRING_LENGTH
145 if(result
<=UCASE_MAX_STRING_LENGTH
) {
147 tmp
.setTo(FALSE
, s
, result
);
148 delta
=result
-U16_LENGTH(c
);
152 delta
=tmp
.length()-U16_LENGTH(c
);
154 text
.handleReplaceBetween(csc
.cpStart
, textPos
, tmp
);
157 csc
.limit
=offsets
.contextLimit
+=delta
;
158 offsets
.limit
+=delta
;
163 offsets
.start
=textPos
;
168 #endif /* #if !UCONFIG_NO_TRANSLITERATION */