1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/ubrk.cpp Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,301 @@
1.4 +/*
1.5 +********************************************************************************
1.6 +* Copyright (C) 1996-2013, International Business Machines
1.7 +* Corporation and others. All Rights Reserved.
1.8 +********************************************************************************
1.9 +*/
1.10 +
1.11 +#include "unicode/utypes.h"
1.12 +
1.13 +#if !UCONFIG_NO_BREAK_ITERATION
1.14 +
1.15 +#include "unicode/ubrk.h"
1.16 +
1.17 +#include "unicode/brkiter.h"
1.18 +#include "unicode/uloc.h"
1.19 +#include "unicode/ustring.h"
1.20 +#include "unicode/uchriter.h"
1.21 +#include "unicode/rbbi.h"
1.22 +#include "rbbirb.h"
1.23 +#include "uassert.h"
1.24 +
1.25 +U_NAMESPACE_USE
1.26 +
1.27 +//------------------------------------------------------------------------------
1.28 +//
1.29 +// ubrk_open Create a canned type of break iterator based on type (word, line, etc.)
1.30 +// and locale.
1.31 +//
1.32 +//------------------------------------------------------------------------------
1.33 +U_CAPI UBreakIterator* U_EXPORT2
1.34 +ubrk_open(UBreakIteratorType type,
1.35 + const char *locale,
1.36 + const UChar *text,
1.37 + int32_t textLength,
1.38 + UErrorCode *status)
1.39 +{
1.40 +
1.41 + if(U_FAILURE(*status)) return 0;
1.42 +
1.43 + BreakIterator *result = 0;
1.44 +
1.45 + switch(type) {
1.46 +
1.47 + case UBRK_CHARACTER:
1.48 + result = BreakIterator::createCharacterInstance(Locale(locale), *status);
1.49 + break;
1.50 +
1.51 + case UBRK_WORD:
1.52 + result = BreakIterator::createWordInstance(Locale(locale), *status);
1.53 + break;
1.54 +
1.55 + case UBRK_LINE:
1.56 + result = BreakIterator::createLineInstance(Locale(locale), *status);
1.57 + break;
1.58 +
1.59 + case UBRK_SENTENCE:
1.60 + result = BreakIterator::createSentenceInstance(Locale(locale), *status);
1.61 + break;
1.62 +
1.63 + case UBRK_TITLE:
1.64 + result = BreakIterator::createTitleInstance(Locale(locale), *status);
1.65 + break;
1.66 +
1.67 + default:
1.68 + *status = U_ILLEGAL_ARGUMENT_ERROR;
1.69 + }
1.70 +
1.71 + // check for allocation error
1.72 + if (U_FAILURE(*status)) {
1.73 + return 0;
1.74 + }
1.75 + if(result == 0) {
1.76 + *status = U_MEMORY_ALLOCATION_ERROR;
1.77 + return 0;
1.78 + }
1.79 +
1.80 +
1.81 + UBreakIterator *uBI = (UBreakIterator *)result;
1.82 + if (text != NULL) {
1.83 + ubrk_setText(uBI, text, textLength, status);
1.84 + }
1.85 + return uBI;
1.86 +}
1.87 +
1.88 +
1.89 +
1.90 +//------------------------------------------------------------------------------
1.91 +//
1.92 +// ubrk_openRules open a break iterator from a set of break rules.
1.93 +// Invokes the rule builder.
1.94 +//
1.95 +//------------------------------------------------------------------------------
1.96 +U_CAPI UBreakIterator* U_EXPORT2
1.97 +ubrk_openRules( const UChar *rules,
1.98 + int32_t rulesLength,
1.99 + const UChar *text,
1.100 + int32_t textLength,
1.101 + UParseError *parseErr,
1.102 + UErrorCode *status) {
1.103 +
1.104 + if (status == NULL || U_FAILURE(*status)){
1.105 + return 0;
1.106 + }
1.107 +
1.108 + BreakIterator *result = 0;
1.109 + UnicodeString ruleString(rules, rulesLength);
1.110 + result = RBBIRuleBuilder::createRuleBasedBreakIterator(ruleString, parseErr, *status);
1.111 + if(U_FAILURE(*status)) {
1.112 + return 0;
1.113 + }
1.114 +
1.115 + UBreakIterator *uBI = (UBreakIterator *)result;
1.116 + if (text != NULL) {
1.117 + ubrk_setText(uBI, text, textLength, status);
1.118 + }
1.119 + return uBI;
1.120 +}
1.121 +
1.122 +
1.123 +
1.124 +
1.125 +
1.126 +U_CAPI UBreakIterator * U_EXPORT2
1.127 +ubrk_safeClone(
1.128 + const UBreakIterator *bi,
1.129 + void * /*stackBuffer*/,
1.130 + int32_t *pBufferSize,
1.131 + UErrorCode *status)
1.132 +{
1.133 + if (status == NULL || U_FAILURE(*status)){
1.134 + return NULL;
1.135 + }
1.136 + if (bi == NULL) {
1.137 + *status = U_ILLEGAL_ARGUMENT_ERROR;
1.138 + return NULL;
1.139 + }
1.140 + if (pBufferSize != NULL) {
1.141 + int32_t inputSize = *pBufferSize;
1.142 + *pBufferSize = 1;
1.143 + if (inputSize == 0) {
1.144 + return NULL; // preflighting for deprecated functionality
1.145 + }
1.146 + }
1.147 + BreakIterator *newBI = ((BreakIterator *)bi)->clone();
1.148 + if (newBI == NULL) {
1.149 + *status = U_MEMORY_ALLOCATION_ERROR;
1.150 + } else {
1.151 + *status = U_SAFECLONE_ALLOCATED_WARNING;
1.152 + }
1.153 + return (UBreakIterator *)newBI;
1.154 +}
1.155 +
1.156 +
1.157 +
1.158 +U_CAPI void U_EXPORT2
1.159 +ubrk_close(UBreakIterator *bi)
1.160 +{
1.161 + delete (BreakIterator *)bi;
1.162 +}
1.163 +
1.164 +U_CAPI void U_EXPORT2
1.165 +ubrk_setText(UBreakIterator* bi,
1.166 + const UChar* text,
1.167 + int32_t textLength,
1.168 + UErrorCode* status)
1.169 +{
1.170 + BreakIterator *brit = (BreakIterator *)bi;
1.171 + UText ut = UTEXT_INITIALIZER;
1.172 + utext_openUChars(&ut, text, textLength, status);
1.173 + brit->setText(&ut, *status);
1.174 + // A stack allocated UText wrapping a UChar * string
1.175 + // can be dumped without explicitly closing it.
1.176 +}
1.177 +
1.178 +
1.179 +
1.180 +U_CAPI void U_EXPORT2
1.181 +ubrk_setUText(UBreakIterator *bi,
1.182 + UText *text,
1.183 + UErrorCode *status)
1.184 +{
1.185 + RuleBasedBreakIterator *brit = (RuleBasedBreakIterator *)bi;
1.186 + brit->RuleBasedBreakIterator::setText(text, *status);
1.187 +}
1.188 +
1.189 +
1.190 +
1.191 +
1.192 +
1.193 +U_CAPI int32_t U_EXPORT2
1.194 +ubrk_current(const UBreakIterator *bi)
1.195 +{
1.196 +
1.197 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::current();
1.198 +}
1.199 +
1.200 +U_CAPI int32_t U_EXPORT2
1.201 +ubrk_next(UBreakIterator *bi)
1.202 +{
1.203 +
1.204 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::next();
1.205 +}
1.206 +
1.207 +U_CAPI int32_t U_EXPORT2
1.208 +ubrk_previous(UBreakIterator *bi)
1.209 +{
1.210 +
1.211 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::previous();
1.212 +}
1.213 +
1.214 +U_CAPI int32_t U_EXPORT2
1.215 +ubrk_first(UBreakIterator *bi)
1.216 +{
1.217 +
1.218 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::first();
1.219 +}
1.220 +
1.221 +U_CAPI int32_t U_EXPORT2
1.222 +ubrk_last(UBreakIterator *bi)
1.223 +{
1.224 +
1.225 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::last();
1.226 +}
1.227 +
1.228 +U_CAPI int32_t U_EXPORT2
1.229 +ubrk_preceding(UBreakIterator *bi,
1.230 + int32_t offset)
1.231 +{
1.232 +
1.233 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::preceding(offset);
1.234 +}
1.235 +
1.236 +U_CAPI int32_t U_EXPORT2
1.237 +ubrk_following(UBreakIterator *bi,
1.238 + int32_t offset)
1.239 +{
1.240 +
1.241 + return ((RuleBasedBreakIterator*)bi)->RuleBasedBreakIterator::following(offset);
1.242 +}
1.243 +
1.244 +U_CAPI const char* U_EXPORT2
1.245 +ubrk_getAvailable(int32_t index)
1.246 +{
1.247 +
1.248 + return uloc_getAvailable(index);
1.249 +}
1.250 +
1.251 +U_CAPI int32_t U_EXPORT2
1.252 +ubrk_countAvailable()
1.253 +{
1.254 +
1.255 + return uloc_countAvailable();
1.256 +}
1.257 +
1.258 +
1.259 +U_CAPI UBool U_EXPORT2
1.260 +ubrk_isBoundary(UBreakIterator *bi, int32_t offset)
1.261 +{
1.262 + return ((RuleBasedBreakIterator *)bi)->RuleBasedBreakIterator::isBoundary(offset);
1.263 +}
1.264 +
1.265 +
1.266 +U_CAPI int32_t U_EXPORT2
1.267 +ubrk_getRuleStatus(UBreakIterator *bi)
1.268 +{
1.269 + return ((RuleBasedBreakIterator *)bi)->RuleBasedBreakIterator::getRuleStatus();
1.270 +}
1.271 +
1.272 +U_CAPI int32_t U_EXPORT2
1.273 +ubrk_getRuleStatusVec(UBreakIterator *bi, int32_t *fillInVec, int32_t capacity, UErrorCode *status)
1.274 +{
1.275 + return ((RuleBasedBreakIterator *)bi)->RuleBasedBreakIterator::getRuleStatusVec(fillInVec, capacity, *status);
1.276 +}
1.277 +
1.278 +
1.279 +U_CAPI const char* U_EXPORT2
1.280 +ubrk_getLocaleByType(const UBreakIterator *bi,
1.281 + ULocDataLocaleType type,
1.282 + UErrorCode* status)
1.283 +{
1.284 + if (bi == NULL) {
1.285 + if (U_SUCCESS(*status)) {
1.286 + *status = U_ILLEGAL_ARGUMENT_ERROR;
1.287 + }
1.288 + return NULL;
1.289 + }
1.290 + return ((BreakIterator*)bi)->getLocaleID(type, *status);
1.291 +}
1.292 +
1.293 +
1.294 +void ubrk_refreshUText(UBreakIterator *bi,
1.295 + UText *text,
1.296 + UErrorCode *status)
1.297 +{
1.298 + BreakIterator *bii = reinterpret_cast<BreakIterator *>(bi);
1.299 + bii->refreshInputText(text, *status);
1.300 +}
1.301 +
1.302 +
1.303 +
1.304 +#endif /* #if !UCONFIG_NO_BREAK_ITERATION */
