1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/brktrans.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,188 @@
1.4 +/*
1.5 +**********************************************************************
1.6 +* Copyright (C) 2008-2010, International Business Machines
1.7 +* Corporation and others. All Rights Reserved.
1.8 +**********************************************************************
1.9 +* Date Name Description
1.10 +* 05/11/2008 Andy Heninger Port from Java
1.11 +**********************************************************************
1.12 +*/
1.13 +
1.14 +#include "unicode/utypes.h"
1.15 +
1.16 +#if !UCONFIG_NO_TRANSLITERATION && !UCONFIG_NO_BREAK_ITERATION
1.17 +
1.18 +#include "unicode/unifilt.h"
1.19 +#include "unicode/uchar.h"
1.20 +#include "unicode/uniset.h"
1.21 +#include "unicode/brkiter.h"
1.22 +#include "brktrans.h"
1.23 +#include "unicode/uchar.h"
1.24 +#include "cmemory.h"
1.25 +#include "uprops.h"
1.26 +#include "uinvchar.h"
1.27 +#include "util.h"
1.28 +#include "uvectr32.h"
1.29 +
1.30 +U_NAMESPACE_BEGIN
1.31 +
1.32 +UOBJECT_DEFINE_RTTI_IMPLEMENTATION(BreakTransliterator)
1.33 +
1.34 +static const UChar SPACE = 32; // ' '
1.35 +
1.36 +
1.37 +/**
1.38 + * Constructs a transliterator with the default delimiters '{' and
1.39 + * '}'.
1.40 + */
1.41 +BreakTransliterator::BreakTransliterator(UnicodeFilter* adoptedFilter) :
1.42 + Transliterator(UNICODE_STRING("Any-BreakInternal", 17), adoptedFilter),
1.43 + fInsertion(SPACE) {
1.44 + bi = NULL;
1.45 + UErrorCode status = U_ZERO_ERROR;
1.46 + boundaries = new UVector32(status);
1.47 + }
1.48 +
1.49 +
1.50 +/**
1.51 + * Destructor.
1.52 + */
1.53 +BreakTransliterator::~BreakTransliterator() {
1.54 + delete bi;
1.55 + bi = NULL;
1.56 + delete boundaries;
1.57 + boundaries = NULL;
1.58 +}
1.59 +
1.60 +/**
1.61 + * Copy constructor.
1.62 + */
1.63 +BreakTransliterator::BreakTransliterator(const BreakTransliterator& o) :
1.64 + Transliterator(o) {
1.65 + bi = NULL;
1.66 + if (o.bi != NULL) {
1.67 + bi = o.bi->clone();
1.68 + }
1.69 + fInsertion = o.fInsertion;
1.70 + UErrorCode status = U_ZERO_ERROR;
1.71 + boundaries = new UVector32(status);
1.72 + }
1.73 +
1.74 +
1.75 +/**
1.76 + * Transliterator API.
1.77 + */
1.78 +Transliterator* BreakTransliterator::clone(void) const {
1.79 + return new BreakTransliterator(*this);
1.80 +}
1.81 +
1.82 +/**
1.83 + * Implements {@link Transliterator#handleTransliterate}.
1.84 + */
1.85 +void BreakTransliterator::handleTransliterate(Replaceable& text, UTransPosition& offsets,
1.86 + UBool isIncremental ) const {
1.87 +
1.88 + UErrorCode status = U_ZERO_ERROR;
1.89 + boundaries->removeAllElements();
1.90 + BreakTransliterator *nonConstThis = (BreakTransliterator *)this;
1.91 + nonConstThis->getBreakIterator(); // Lazy-create it if necessary
1.92 + UnicodeString sText = replaceableAsString(text);
1.93 + bi->setText(sText);
1.94 + bi->preceding(offsets.start);
1.95 +
1.96 + // To make things much easier, we will stack the boundaries, and then insert at the end.
1.97 + // generally, we won't need too many, since we will be filtered.
1.98 +
1.99 + int32_t boundary;
1.100 + for(boundary = bi->next(); boundary != UBRK_DONE && boundary < offsets.limit; boundary = bi->next()) {
1.101 + if (boundary == 0) continue;
1.102 + // HACK: Check to see that preceeding item was a letter
1.103 +
1.104 + UChar32 cp = sText.char32At(boundary-1);
1.105 + int type = u_charType(cp);
1.106 + //System.out.println(Integer.toString(cp,16) + " (before): " + type);
1.107 + if ((U_MASK(type) & (U_GC_L_MASK | U_GC_M_MASK)) == 0) continue;
1.108 +
1.109 + cp = sText.char32At(boundary);
1.110 + type = u_charType(cp);
1.111 + //System.out.println(Integer.toString(cp,16) + " (after): " + type);
1.112 + if ((U_MASK(type) & (U_GC_L_MASK | U_GC_M_MASK)) == 0) continue;
1.113 +
1.114 + boundaries->addElement(boundary, status);
1.115 + // printf("Boundary at %d\n", boundary);
1.116 + }
1.117 +
1.118 + int delta = 0;
1.119 + int lastBoundary = 0;
1.120 +
1.121 + if (boundaries->size() != 0) { // if we found something, adjust
1.122 + delta = boundaries->size() * fInsertion.length();
1.123 + lastBoundary = boundaries->lastElementi();
1.124 +
1.125 + // we do this from the end backwards, so that we don't have to keep updating.
1.126 +
1.127 + while (boundaries->size() > 0) {
1.128 + boundary = boundaries->popi();
1.129 + text.handleReplaceBetween(boundary, boundary, fInsertion);
1.130 + }
1.131 + }
1.132 +
1.133 + // Now fix up the return values
1.134 + offsets.contextLimit += delta;
1.135 + offsets.limit += delta;
1.136 + offsets.start = isIncremental ? lastBoundary + delta : offsets.limit;
1.137 +
1.138 + // TODO: do something with U_FAILURE(status);
1.139 + // (need to look at transliterators overall, not just here.)
1.140 +}
1.141 +
1.142 +//
1.143 +// getInsertion()
1.144 +//
1.145 +const UnicodeString &BreakTransliterator::getInsertion() const {
1.146 + return fInsertion;
1.147 +}
1.148 +
1.149 +//
1.150 +// setInsertion()
1.151 +//
1.152 +void BreakTransliterator::setInsertion(const UnicodeString &insertion) {
1.153 + this->fInsertion = insertion;
1.154 +}
1.155 +
1.156 +//
1.157 +// getBreakIterator Lazily create the break iterator if it does
1.158 +// not already exist. Copied from Java, probably
1.159 +// better to just create it in the constructor.
1.160 +//
1.161 +BreakIterator *BreakTransliterator::getBreakIterator() {
1.162 + UErrorCode status = U_ZERO_ERROR;
1.163 + if (bi == NULL) {
1.164 + // Note: Thai breaking behavior is universal, it is not
1.165 + // tied to the Thai locale.
1.166 + bi = BreakIterator::createWordInstance(Locale::getEnglish(), status);
1.167 + }
1.168 + return bi;
1.169 +}
1.170 +
1.171 +//
1.172 +// replaceableAsString Hack to let break iterators work
1.173 +// on the replaceable text from transliterators.
1.174 +// In practice, the only real Replaceable type that we
1.175 +// will be seeing is UnicodeString, so this function
1.176 +// will normally be efficient.
1.177 +//
1.178 +UnicodeString BreakTransliterator::replaceableAsString(Replaceable &r) {
1.179 + UnicodeString s;
1.180 + UnicodeString *rs = dynamic_cast<UnicodeString *>(&r);
1.181 + if (rs != NULL) {
1.182 + s = *rs;
1.183 + } else {
1.184 + r.extractBetween(0, r.length(), s);
1.185 + }
1.186 + return s;
1.187 +}
1.188 +
1.189 +U_NAMESPACE_END
1.190 +
1.191 +#endif /* #if !UCONFIG_NO_TRANSLITERATION */
