1 // © 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
3 /*
4 *******************************************************************************
5 * Copyright (C) 2011, International Business Machines
6 * Corporation and others. All Rights Reserved.
7 *******************************************************************************
8 * file name: ustr_titlecase_brkiter.cpp
9 * encoding: UTF-8
10 * tab size: 8 (not used)
11 * indentation:4
12 *
13 * created on: 2011may30
14 * created by: Markus W. Scherer
15 *
16 * Titlecasing functions that are based on BreakIterator
17 * were moved here to break dependency cycles among parts of the common library.
18 */
19
20 #include "unicode/utypes.h"
21
22 #if !UCONFIG_NO_BREAK_ITERATION
23
24 #include "unicode/brkiter.h"
25 #include "unicode/casemap.h"
26 #include "unicode/chariter.h"
27 #include "unicode/localpointer.h"
28 #include "unicode/ubrk.h"
29 #include "unicode/ucasemap.h"
30 #include "unicode/utext.h"
31 #include "cmemory.h"
32 #include "uassert.h"
33 #include "ucase.h"
34 #include "ucasemap_imp.h"
35
36 U_NAMESPACE_BEGIN
37
38 /**
39 * Whole-string BreakIterator.
40 * Titlecasing only calls setText(), first(), and next().
41 * We implement the rest only to satisfy the abstract interface.
42 */
43 class WholeStringBreakIterator : public BreakIterator {
44 public:
WholeStringBreakIterator()45 WholeStringBreakIterator() : BreakIterator(), length(0) {}
46 ~WholeStringBreakIterator() U_OVERRIDE;
47 bool operator==(const BreakIterator&) const U_OVERRIDE;
48 WholeStringBreakIterator *clone() const U_OVERRIDE;
49 static UClassID U_EXPORT2 getStaticClassID();
50 UClassID getDynamicClassID() const U_OVERRIDE;
51 CharacterIterator &getText() const U_OVERRIDE;
52 UText *getUText(UText *fillIn, UErrorCode &errorCode) const U_OVERRIDE;
53 void setText(const UnicodeString &text) U_OVERRIDE;
54 void setText(UText *text, UErrorCode &errorCode) U_OVERRIDE;
55 void adoptText(CharacterIterator* it) U_OVERRIDE;
56 int32_t first() U_OVERRIDE;
57 int32_t last() U_OVERRIDE;
58 int32_t previous() U_OVERRIDE;
59 int32_t next() U_OVERRIDE;
60 int32_t current() const U_OVERRIDE;
61 int32_t following(int32_t offset) U_OVERRIDE;
62 int32_t preceding(int32_t offset) U_OVERRIDE;
63 UBool isBoundary(int32_t offset) U_OVERRIDE;
64 int32_t next(int32_t n) U_OVERRIDE;
65 WholeStringBreakIterator *createBufferClone(void *stackBuffer, int32_t &BufferSize,
66 UErrorCode &errorCode) U_OVERRIDE;
67 WholeStringBreakIterator &refreshInputText(UText *input, UErrorCode &errorCode) U_OVERRIDE;
68
69 private:
70 int32_t length;
71 };
72
UOBJECT_DEFINE_RTTI_IMPLEMENTATION(WholeStringBreakIterator)73 UOBJECT_DEFINE_RTTI_IMPLEMENTATION(WholeStringBreakIterator)
74
75 WholeStringBreakIterator::~WholeStringBreakIterator() {}
operator ==(const BreakIterator &) const76 bool WholeStringBreakIterator::operator==(const BreakIterator&) const { return false; }
clone() const77 WholeStringBreakIterator *WholeStringBreakIterator::clone() const { return nullptr; }
78
getText() const79 CharacterIterator &WholeStringBreakIterator::getText() const {
80 UPRV_UNREACHABLE_EXIT; // really should not be called
81 }
getUText(UText *,UErrorCode & errorCode) const82 UText *WholeStringBreakIterator::getUText(UText * /*fillIn*/, UErrorCode &errorCode) const {
83 if (U_SUCCESS(errorCode)) {
84 errorCode = U_UNSUPPORTED_ERROR;
85 }
86 return nullptr;
87 }
88
setText(const UnicodeString & text)89 void WholeStringBreakIterator::setText(const UnicodeString &text) {
90 length = text.length();
91 }
setText(UText * text,UErrorCode & errorCode)92 void WholeStringBreakIterator::setText(UText *text, UErrorCode &errorCode) {
93 if (U_SUCCESS(errorCode)) {
94 int64_t length64 = utext_nativeLength(text);
95 if (length64 <= INT32_MAX) {
96 length = (int32_t)length64;
97 } else {
98 errorCode = U_INDEX_OUTOFBOUNDS_ERROR;
99 }
100 }
101 }
adoptText(CharacterIterator *)102 void WholeStringBreakIterator::adoptText(CharacterIterator*) {
103 UPRV_UNREACHABLE_EXIT; // should not be called
104 }
105
first()106 int32_t WholeStringBreakIterator::first() { return 0; }
last()107 int32_t WholeStringBreakIterator::last() { return length; }
previous()108 int32_t WholeStringBreakIterator::previous() { return 0; }
next()109 int32_t WholeStringBreakIterator::next() { return length; }
current() const110 int32_t WholeStringBreakIterator::current() const { return 0; }
following(int32_t)111 int32_t WholeStringBreakIterator::following(int32_t /*offset*/) { return length; }
preceding(int32_t)112 int32_t WholeStringBreakIterator::preceding(int32_t /*offset*/) { return 0; }
isBoundary(int32_t)113 UBool WholeStringBreakIterator::isBoundary(int32_t /*offset*/) { return FALSE; }
next(int32_t)114 int32_t WholeStringBreakIterator::next(int32_t /*n*/) { return length; }
115
createBufferClone(void *,int32_t &,UErrorCode & errorCode)116 WholeStringBreakIterator *WholeStringBreakIterator::createBufferClone(
117 void * /*stackBuffer*/, int32_t & /*BufferSize*/, UErrorCode &errorCode) {
118 if (U_SUCCESS(errorCode)) {
119 errorCode = U_UNSUPPORTED_ERROR;
120 }
121 return nullptr;
122 }
refreshInputText(UText *,UErrorCode & errorCode)123 WholeStringBreakIterator &WholeStringBreakIterator::refreshInputText(
124 UText * /*input*/, UErrorCode &errorCode) {
125 if (U_SUCCESS(errorCode)) {
126 errorCode = U_UNSUPPORTED_ERROR;
127 }
128 return *this;
129 }
130
131 U_CFUNC
ustrcase_getTitleBreakIterator(const Locale * locale,const char * locID,uint32_t options,BreakIterator * iter,LocalPointer<BreakIterator> & ownedIter,UErrorCode & errorCode)132 BreakIterator *ustrcase_getTitleBreakIterator(
133 const Locale *locale, const char *locID, uint32_t options, BreakIterator *iter,
134 LocalPointer<BreakIterator> &ownedIter, UErrorCode &errorCode) {
135 if (U_FAILURE(errorCode)) { return nullptr; }
136 options &= U_TITLECASE_ITERATOR_MASK;
137 if (options != 0 && iter != nullptr) {
138 errorCode = U_ILLEGAL_ARGUMENT_ERROR;
139 return nullptr;
140 }
141 if (iter == nullptr) {
142 switch (options) {
143 case 0:
144 iter = BreakIterator::createWordInstance(
145 locale != nullptr ? *locale : Locale(locID), errorCode);
146 break;
147 case U_TITLECASE_WHOLE_STRING:
148 iter = new WholeStringBreakIterator();
149 if (iter == nullptr) {
150 errorCode = U_MEMORY_ALLOCATION_ERROR;
151 }
152 break;
153 case U_TITLECASE_SENTENCES:
154 iter = BreakIterator::createSentenceInstance(
155 locale != nullptr ? *locale : Locale(locID), errorCode);
156 break;
157 default:
158 errorCode = U_ILLEGAL_ARGUMENT_ERROR;
159 break;
160 }
161 ownedIter.adoptInstead(iter);
162 }
163 return iter;
164 }
165
toTitle(const char * locale,uint32_t options,BreakIterator * iter,const UChar * src,int32_t srcLength,UChar * dest,int32_t destCapacity,Edits * edits,UErrorCode & errorCode)166 int32_t CaseMap::toTitle(
167 const char *locale, uint32_t options, BreakIterator *iter,
168 const UChar *src, int32_t srcLength,
169 UChar *dest, int32_t destCapacity, Edits *edits,
170 UErrorCode &errorCode) {
171 LocalPointer<BreakIterator> ownedIter;
172 iter = ustrcase_getTitleBreakIterator(nullptr, locale, options, iter, ownedIter, errorCode);
173 if(iter==NULL) {
174 return 0;
175 }
176 UnicodeString s(srcLength<0, src, srcLength);
177 iter->setText(s);
178 return ustrcase_map(
179 ustrcase_getCaseLocale(locale), options, iter,
180 dest, destCapacity,
181 src, srcLength,
182 ustrcase_internalToTitle, edits, errorCode);
183 }
184
185 U_NAMESPACE_END
186
187 U_NAMESPACE_USE
188
189 U_CAPI int32_t U_EXPORT2
u_strToTitle(UChar * dest,int32_t destCapacity,const UChar * src,int32_t srcLength,UBreakIterator * titleIter,const char * locale,UErrorCode * pErrorCode)190 u_strToTitle(UChar *dest, int32_t destCapacity,
191 const UChar *src, int32_t srcLength,
192 UBreakIterator *titleIter,
193 const char *locale,
194 UErrorCode *pErrorCode) {
195 LocalPointer<BreakIterator> ownedIter;
196 BreakIterator *iter = ustrcase_getTitleBreakIterator(
197 nullptr, locale, 0, reinterpret_cast<BreakIterator *>(titleIter),
198 ownedIter, *pErrorCode);
199 if (iter == nullptr) {
200 return 0;
201 }
202 UnicodeString s(srcLength<0, src, srcLength);
203 iter->setText(s);
204 return ustrcase_mapWithOverlap(
205 ustrcase_getCaseLocale(locale), 0, iter,
206 dest, destCapacity,
207 src, srcLength,
208 ustrcase_internalToTitle, *pErrorCode);
209 }
210
211 U_CAPI int32_t U_EXPORT2
ucasemap_toTitle(UCaseMap * csm,UChar * dest,int32_t destCapacity,const UChar * src,int32_t srcLength,UErrorCode * pErrorCode)212 ucasemap_toTitle(UCaseMap *csm,
213 UChar *dest, int32_t destCapacity,
214 const UChar *src, int32_t srcLength,
215 UErrorCode *pErrorCode) {
216 if (U_FAILURE(*pErrorCode)) {
217 return 0;
218 }
219 if (csm->iter == NULL) {
220 LocalPointer<BreakIterator> ownedIter;
221 BreakIterator *iter = ustrcase_getTitleBreakIterator(
222 nullptr, csm->locale, csm->options, nullptr, ownedIter, *pErrorCode);
223 if (iter == nullptr) {
224 return 0;
225 }
226 csm->iter = ownedIter.orphan();
227 }
228 UnicodeString s(srcLength<0, src, srcLength);
229 csm->iter->setText(s);
230 return ustrcase_map(
231 csm->caseLocale, csm->options, csm->iter,
232 dest, destCapacity,
233 src, srcLength,
234 ustrcase_internalToTitle, NULL, *pErrorCode);
235 }
236
237 #endif // !UCONFIG_NO_BREAK_ITERATION
238