1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/titletrn.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,168 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (C) 2001-2011, International Business Machines
1.7 +* Corporation and others. All Rights Reserved.
1.8 +**********************************************************************
1.9 +* Date Name Description
1.10 +* 05/24/01 aliu Creation.
1.11 +**********************************************************************
1.12 +*/
1.13 +
1.14 +#include "unicode/utypes.h"
1.15 +
1.16 +#if !UCONFIG_NO_TRANSLITERATION
1.17 +
1.18 +#include "unicode/uchar.h"
1.19 +#include "unicode/uniset.h"
1.20 +#include "unicode/ustring.h"
1.21 +#include "unicode/utf16.h"
1.22 +#include "titletrn.h"
1.23 +#include "umutex.h"
1.24 +#include "ucase.h"
1.25 +#include "cpputils.h"
1.26 +
1.27 +U_NAMESPACE_BEGIN
1.28 +
1.29 +UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TitlecaseTransliterator)
1.30 +
1.31 +TitlecaseTransliterator::TitlecaseTransliterator() :
1.32 + CaseMapTransliterator(UNICODE_STRING("Any-Title", 9), NULL)
1.33 +{
1.34 + // Need to look back 2 characters in the case of "can't"
1.35 + setMaximumContextLength(2);
1.36 +}
1.37 +
1.38 +/**
1.39 + * Destructor.
1.40 + */
1.41 +TitlecaseTransliterator::~TitlecaseTransliterator() {
1.42 +}
1.43 +
1.44 +/**
1.45 + * Copy constructor.
1.46 + */
1.47 +TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator& o) :
1.48 + CaseMapTransliterator(o)
1.49 +{
1.50 +}
1.51 +
1.52 +/**
1.53 + * Assignment operator.
1.54 + */
1.55 +/*TitlecaseTransliterator& TitlecaseTransliterator::operator=(
1.56 + const TitlecaseTransliterator& o) {
1.57 + CaseMapTransliterator::operator=(o);
1.58 + return *this;
1.59 +}*/
1.60 +
1.61 +/**
1.62 + * Transliterator API.
1.63 + */
1.64 +Transliterator* TitlecaseTransliterator::clone(void) const {
1.65 + return new TitlecaseTransliterator(*this);
1.66 +}
1.67 +
1.68 +/**
1.69 + * Implements {@link Transliterator#handleTransliterate}.
1.70 + */
1.71 +void TitlecaseTransliterator::handleTransliterate(
1.72 + Replaceable& text, UTransPosition& offsets,
1.73 + UBool isIncremental) const
1.74 +{
1.75 + // TODO reimplement, see ustrcase.c
1.76 + // using a real word break iterator
1.77 + // instead of just looking for a transition between cased and uncased characters
1.78 + // call CaseMapTransliterator::handleTransliterate() for lowercasing? (set fMap)
1.79 + // needs to take isIncremental into account because case mappings are context-sensitive
1.80 + // also detect when lowercasing function did not finish because of context
1.81 +
1.82 + if (offsets.start >= offsets.limit) {
1.83 + return;
1.84 + }
1.85 +
1.86 + // case type: >0 cased (UCASE_LOWER etc.) ==0 uncased <0 case-ignorable
1.87 + int32_t type;
1.88 +
1.89 + // Our mode; we are either converting letter toTitle or
1.90 + // toLower.
1.91 + UBool doTitle = TRUE;
1.92 +
1.93 + // Determine if there is a preceding context of cased case-ignorable*,
1.94 + // in which case we want to start in toLower mode. If the
1.95 + // prior context is anything else (including empty) then start
1.96 + // in toTitle mode.
1.97 + UChar32 c;
1.98 + int32_t start;
1.99 + for (start = offsets.start - 1; start >= offsets.contextStart; start -= U16_LENGTH(c)) {
1.100 + c = text.char32At(start);
1.101 + type=ucase_getTypeOrIgnorable(fCsp, c);
1.102 + if(type>0) { // cased
1.103 + doTitle=FALSE;
1.104 + break;
1.105 + } else if(type==0) { // uncased but not ignorable
1.106 + break;
1.107 + }
1.108 + // else (type<0) case-ignorable: continue
1.109 + }
1.110 +
1.111 + // Convert things after a cased character toLower; things
1.112 + // after an uncased, non-case-ignorable character toTitle. Case-ignorable
1.113 + // characters are copied directly and do not change the mode.
1.114 + UCaseContext csc;
1.115 + uprv_memset(&csc, 0, sizeof(csc));
1.116 + csc.p = &text;
1.117 + csc.start = offsets.contextStart;
1.118 + csc.limit = offsets.contextLimit;
1.119 +
1.120 + UnicodeString tmp;
1.121 + const UChar *s;
1.122 + int32_t textPos, delta, result, locCache=0;
1.123 +
1.124 + for(textPos=offsets.start; textPos<offsets.limit;) {
1.125 + csc.cpStart=textPos;
1.126 + c=text.char32At(textPos);
1.127 + csc.cpLimit=textPos+=U16_LENGTH(c);
1.128 +
1.129 + type=ucase_getTypeOrIgnorable(fCsp, c);
1.130 + if(type>=0) { // not case-ignorable
1.131 + if(doTitle) {
1.132 + result=ucase_toFullTitle(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache);
1.133 + } else {
1.134 + result=ucase_toFullLower(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache);
1.135 + }
1.136 + doTitle = (UBool)(type==0); // doTitle=isUncased
1.137 +
1.138 + if(csc.b1 && isIncremental) {
1.139 + // fMap() tried to look beyond the context limit
1.140 + // wait for more input
1.141 + offsets.start=csc.cpStart;
1.142 + return;
1.143 + }
1.144 +
1.145 + if(result>=0) {
1.146 + // replace the current code point with its full case mapping result
1.147 + // see UCASE_MAX_STRING_LENGTH
1.148 + if(result<=UCASE_MAX_STRING_LENGTH) {
1.149 + // string s[result]
1.150 + tmp.setTo(FALSE, s, result);
1.151 + delta=result-U16_LENGTH(c);
1.152 + } else {
1.153 + // single code point
1.154 + tmp.setTo(result);
1.155 + delta=tmp.length()-U16_LENGTH(c);
1.156 + }
1.157 + text.handleReplaceBetween(csc.cpStart, textPos, tmp);
1.158 + if(delta!=0) {
1.159 + textPos+=delta;
1.160 + csc.limit=offsets.contextLimit+=delta;
1.161 + offsets.limit+=delta;
1.162 + }
1.163 + }
1.164 + }
1.165 + }
1.166 + offsets.start=textPos;
1.167 +}
1.168 +
1.169 +U_NAMESPACE_END
1.170 +
1.171 +#endif /* #if !UCONFIG_NO_TRANSLITERATION */
