]>
Commit | Line | Data |
---|---|---|
1 | /* | |
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 | ********************************************************************** | |
9 | */ | |
10 | ||
11 | #include "unicode/utypes.h" | |
12 | ||
13 | #if !UCONFIG_NO_TRANSLITERATION | |
14 | ||
15 | #include "unicode/uchar.h" | |
16 | #include "unicode/uniset.h" | |
17 | #include "unicode/ustring.h" | |
18 | #include "unicode/utf16.h" | |
19 | #include "titletrn.h" | |
20 | #include "umutex.h" | |
21 | #include "ucase.h" | |
22 | #include "cpputils.h" | |
23 | ||
24 | U_NAMESPACE_BEGIN | |
25 | ||
26 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TitlecaseTransliterator) | |
27 | ||
28 | TitlecaseTransliterator::TitlecaseTransliterator() : | |
29 | CaseMapTransliterator(UNICODE_STRING("Any-Title", 9), NULL) | |
30 | { | |
31 | // Need to look back 2 characters in the case of "can't" | |
32 | setMaximumContextLength(2); | |
33 | } | |
34 | ||
35 | /** | |
36 | * Destructor. | |
37 | */ | |
38 | TitlecaseTransliterator::~TitlecaseTransliterator() { | |
39 | } | |
40 | ||
41 | /** | |
42 | * Copy constructor. | |
43 | */ | |
44 | TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator& o) : | |
45 | CaseMapTransliterator(o) | |
46 | { | |
47 | } | |
48 | ||
49 | /** | |
50 | * Assignment operator. | |
51 | */ | |
52 | /*TitlecaseTransliterator& TitlecaseTransliterator::operator=( | |
53 | const TitlecaseTransliterator& o) { | |
54 | CaseMapTransliterator::operator=(o); | |
55 | return *this; | |
56 | }*/ | |
57 | ||
58 | /** | |
59 | * Transliterator API. | |
60 | */ | |
61 | Transliterator* TitlecaseTransliterator::clone(void) const { | |
62 | return new TitlecaseTransliterator(*this); | |
63 | } | |
64 | ||
65 | /** | |
66 | * Implements {@link Transliterator#handleTransliterate}. | |
67 | */ | |
68 | void TitlecaseTransliterator::handleTransliterate( | |
69 | Replaceable& text, UTransPosition& offsets, | |
70 | UBool isIncremental) const | |
71 | { | |
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 | |
78 | ||
79 | if (offsets.start >= offsets.limit) { | |
80 | return; | |
81 | } | |
82 | ||
83 | // case type: >0 cased (UCASE_LOWER etc.) ==0 uncased <0 case-ignorable | |
84 | int32_t type; | |
85 | ||
86 | // Our mode; we are either converting letter toTitle or | |
87 | // toLower. | |
88 | UBool doTitle = TRUE; | |
89 | ||
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 | |
93 | // in toTitle mode. | |
94 | UChar32 c; | |
95 | int32_t 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); | |
99 | if(type>0) { // cased | |
100 | doTitle=FALSE; | |
101 | break; | |
102 | } else if(type==0) { // uncased but not ignorable | |
103 | break; | |
104 | } | |
105 | // else (type<0) case-ignorable: continue | |
106 | } | |
107 | ||
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. | |
111 | UCaseContext csc; | |
112 | uprv_memset(&csc, 0, sizeof(csc)); | |
113 | csc.p = &text; | |
114 | csc.start = offsets.contextStart; | |
115 | csc.limit = offsets.contextLimit; | |
116 | ||
117 | UnicodeString tmp; | |
118 | const UChar *s; | |
119 | int32_t textPos, delta, result, locCache=0; | |
120 | ||
121 | for(textPos=offsets.start; textPos<offsets.limit;) { | |
122 | csc.cpStart=textPos; | |
123 | c=text.char32At(textPos); | |
124 | csc.cpLimit=textPos+=U16_LENGTH(c); | |
125 | ||
126 | type=ucase_getTypeOrIgnorable(fCsp, c); | |
127 | if(type>=0) { // not case-ignorable | |
128 | if(doTitle) { | |
129 | result=ucase_toFullTitle(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache); | |
130 | } else { | |
131 | result=ucase_toFullLower(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache); | |
132 | } | |
133 | doTitle = (UBool)(type==0); // doTitle=isUncased | |
134 | ||
135 | if(csc.b1 && isIncremental) { | |
136 | // fMap() tried to look beyond the context limit | |
137 | // wait for more input | |
138 | offsets.start=csc.cpStart; | |
139 | return; | |
140 | } | |
141 | ||
142 | if(result>=0) { | |
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) { | |
146 | // string s[result] | |
147 | tmp.setTo(FALSE, s, result); | |
148 | delta=result-U16_LENGTH(c); | |
149 | } else { | |
150 | // single code point | |
151 | tmp.setTo(result); | |
152 | delta=tmp.length()-U16_LENGTH(c); | |
153 | } | |
154 | text.handleReplaceBetween(csc.cpStart, textPos, tmp); | |
155 | if(delta!=0) { | |
156 | textPos+=delta; | |
157 | csc.limit=offsets.contextLimit+=delta; | |
158 | offsets.limit+=delta; | |
159 | } | |
160 | } | |
161 | } | |
162 | } | |
163 | offsets.start=textPos; | |
164 | } | |
165 | ||
166 | U_NAMESPACE_END | |
167 | ||
168 | #endif /* #if !UCONFIG_NO_TRANSLITERATION */ |