1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/casetrn.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,193 @@
1.4 +/*
1.5 +*******************************************************************************
1.6 +*
1.7 +* Copyright (C) 2001-2011, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +*******************************************************************************
1.11 +* file name: casetrn.cpp
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2004sep03
1.17 +* created by: Markus W. Scherer
1.18 +*
1.19 +* Implementation class for lower-/upper-/title-casing transliterators.
1.20 +*/
1.21 +
1.22 +#include "unicode/utypes.h"
1.23 +
1.24 +#if !UCONFIG_NO_TRANSLITERATION
1.25 +
1.26 +#include "unicode/uchar.h"
1.27 +#include "unicode/ustring.h"
1.28 +#include "unicode/utf.h"
1.29 +#include "unicode/utf16.h"
1.30 +#include "tolowtrn.h"
1.31 +#include "ucase.h"
1.32 +#include "cpputils.h"
1.33 +
1.34 +/* case context iterator using a Replaceable */
1.35 +U_CFUNC UChar32 U_CALLCONV
1.36 +utrans_rep_caseContextIterator(void *context, int8_t dir)
1.37 +{
1.38 + U_NAMESPACE_USE
1.39 +
1.40 + UCaseContext *csc=(UCaseContext *)context;
1.41 + Replaceable *rep=(Replaceable *)csc->p;
1.42 + UChar32 c;
1.43 +
1.44 + if(dir<0) {
1.45 + /* reset for backward iteration */
1.46 + csc->index=csc->cpStart;
1.47 + csc->dir=dir;
1.48 + } else if(dir>0) {
1.49 + /* reset for forward iteration */
1.50 + csc->index=csc->cpLimit;
1.51 + csc->dir=dir;
1.52 + } else {
1.53 + /* continue current iteration direction */
1.54 + dir=csc->dir;
1.55 + }
1.56 +
1.57 + // automatically adjust start and limit if the Replaceable disagrees
1.58 + // with the original values
1.59 + if(dir<0) {
1.60 + if(csc->start<csc->index) {
1.61 + c=rep->char32At(csc->index-1);
1.62 + if(c<0) {
1.63 + csc->start=csc->index;
1.64 + } else {
1.65 + csc->index-=U16_LENGTH(c);
1.66 + return c;
1.67 + }
1.68 + }
1.69 + } else {
1.70 + // detect, and store in csc->b1, if we hit the limit
1.71 + if(csc->index<csc->limit) {
1.72 + c=rep->char32At(csc->index);
1.73 + if(c<0) {
1.74 + csc->limit=csc->index;
1.75 + csc->b1=TRUE;
1.76 + } else {
1.77 + csc->index+=U16_LENGTH(c);
1.78 + return c;
1.79 + }
1.80 + } else {
1.81 + csc->b1=TRUE;
1.82 + }
1.83 + }
1.84 + return U_SENTINEL;
1.85 +}
1.86 +
1.87 +U_NAMESPACE_BEGIN
1.88 +
1.89 +UOBJECT_DEFINE_ABSTRACT_RTTI_IMPLEMENTATION(CaseMapTransliterator)
1.90 +
1.91 +/**
1.92 + * Constructs a transliterator.
1.93 + */
1.94 +CaseMapTransliterator::CaseMapTransliterator(const UnicodeString &id, UCaseMapFull *map) :
1.95 + Transliterator(id, 0),
1.96 + fCsp(ucase_getSingleton()),
1.97 + fMap(map)
1.98 +{
1.99 + // TODO test incremental mode with context-sensitive text (e.g. greek sigma)
1.100 + // TODO need to call setMaximumContextLength()?!
1.101 +}
1.102 +
1.103 +/**
1.104 + * Destructor.
1.105 + */
1.106 +CaseMapTransliterator::~CaseMapTransliterator() {
1.107 +}
1.108 +
1.109 +/**
1.110 + * Copy constructor.
1.111 + */
1.112 +CaseMapTransliterator::CaseMapTransliterator(const CaseMapTransliterator& o) :
1.113 + Transliterator(o),
1.114 + fCsp(o.fCsp), fMap(o.fMap)
1.115 +{
1.116 +}
1.117 +
1.118 +/**
1.119 + * Assignment operator.
1.120 + */
1.121 +/*CaseMapTransliterator& CaseMapTransliterator::operator=(const CaseMapTransliterator& o) {
1.122 + Transliterator::operator=(o);
1.123 + fCsp = o.fCsp;
1.124 + fMap = o.fMap;
1.125 + return *this;
1.126 +}*/
1.127 +
1.128 +/**
1.129 + * Transliterator API.
1.130 + */
1.131 +/*Transliterator* CaseMapTransliterator::clone(void) const {
1.132 + return new CaseMapTransliterator(*this);
1.133 +}*/
1.134 +
1.135 +/**
1.136 + * Implements {@link Transliterator#handleTransliterate}.
1.137 + */
1.138 +void CaseMapTransliterator::handleTransliterate(Replaceable& text,
1.139 + UTransPosition& offsets,
1.140 + UBool isIncremental) const
1.141 +{
1.142 + if (offsets.start >= offsets.limit) {
1.143 + return;
1.144 + }
1.145 +
1.146 + UCaseContext csc;
1.147 + uprv_memset(&csc, 0, sizeof(csc));
1.148 + csc.p = &text;
1.149 + csc.start = offsets.contextStart;
1.150 + csc.limit = offsets.contextLimit;
1.151 +
1.152 + UnicodeString tmp;
1.153 + const UChar *s;
1.154 + UChar32 c;
1.155 + int32_t textPos, delta, result, locCache=0;
1.156 +
1.157 + for(textPos=offsets.start; textPos<offsets.limit;) {
1.158 + csc.cpStart=textPos;
1.159 + c=text.char32At(textPos);
1.160 + csc.cpLimit=textPos+=U16_LENGTH(c);
1.161 +
1.162 + result=fMap(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache);
1.163 +
1.164 + if(csc.b1 && isIncremental) {
1.165 + // fMap() tried to look beyond the context limit
1.166 + // wait for more input
1.167 + offsets.start=csc.cpStart;
1.168 + return;
1.169 + }
1.170 +
1.171 + if(result>=0) {
1.172 + // replace the current code point with its full case mapping result
1.173 + // see UCASE_MAX_STRING_LENGTH
1.174 + if(result<=UCASE_MAX_STRING_LENGTH) {
1.175 + // string s[result]
1.176 + tmp.setTo(FALSE, s, result);
1.177 + delta=result-U16_LENGTH(c);
1.178 + } else {
1.179 + // single code point
1.180 + tmp.setTo(result);
1.181 + delta=tmp.length()-U16_LENGTH(c);
1.182 + }
1.183 + text.handleReplaceBetween(csc.cpStart, textPos, tmp);
1.184 + if(delta!=0) {
1.185 + textPos+=delta;
1.186 + csc.limit=offsets.contextLimit+=delta;
1.187 + offsets.limit+=delta;
1.188 + }
1.189 + }
1.190 + }
1.191 + offsets.start=textPos;
1.192 +}
1.193 +
1.194 +U_NAMESPACE_END
1.195 +
1.196 +#endif /* #if !UCONFIG_NO_TRANSLITERATION */
