michael@0: /*
michael@0: **********************************************************************
michael@0: *   Copyright (C) 2001-2011, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: **********************************************************************
michael@0: *   Date        Name        Description
michael@0: *   05/24/01    aliu        Creation.
michael@0: **********************************************************************
michael@0: */
michael@0: 
michael@0: #include "unicode/utypes.h"
michael@0: 
michael@0: #if !UCONFIG_NO_TRANSLITERATION
michael@0: 
michael@0: #include "unicode/uchar.h"
michael@0: #include "unicode/uniset.h"
michael@0: #include "unicode/ustring.h"
michael@0: #include "unicode/utf16.h"
michael@0: #include "titletrn.h"
michael@0: #include "umutex.h"
michael@0: #include "ucase.h"
michael@0: #include "cpputils.h"
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TitlecaseTransliterator)
michael@0: 
michael@0: TitlecaseTransliterator::TitlecaseTransliterator() :
michael@0:     CaseMapTransliterator(UNICODE_STRING("Any-Title", 9), NULL)
michael@0: {
michael@0:     // Need to look back 2 characters in the case of "can't"
michael@0:     setMaximumContextLength(2);
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Destructor.
michael@0:  */
michael@0: TitlecaseTransliterator::~TitlecaseTransliterator() {
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Copy constructor.
michael@0:  */
michael@0: TitlecaseTransliterator::TitlecaseTransliterator(const TitlecaseTransliterator& o) :
michael@0:     CaseMapTransliterator(o)
michael@0: {
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Assignment operator.
michael@0:  */
michael@0: /*TitlecaseTransliterator& TitlecaseTransliterator::operator=(
michael@0:                              const TitlecaseTransliterator& o) {
michael@0:     CaseMapTransliterator::operator=(o);
michael@0:     return *this;
michael@0: }*/
michael@0: 
michael@0: /**
michael@0:  * Transliterator API.
michael@0:  */
michael@0: Transliterator* TitlecaseTransliterator::clone(void) const {
michael@0:     return new TitlecaseTransliterator(*this);
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Implements {@link Transliterator#handleTransliterate}.
michael@0:  */
michael@0: void TitlecaseTransliterator::handleTransliterate(
michael@0:                                   Replaceable& text, UTransPosition& offsets,
michael@0:                                   UBool isIncremental) const
michael@0: {
michael@0:     // TODO reimplement, see ustrcase.c
michael@0:     // using a real word break iterator
michael@0:     //   instead of just looking for a transition between cased and uncased characters
michael@0:     // call CaseMapTransliterator::handleTransliterate() for lowercasing? (set fMap)
michael@0:     // needs to take isIncremental into account because case mappings are context-sensitive
michael@0:     //   also detect when lowercasing function did not finish because of context
michael@0: 
michael@0:     if (offsets.start >= offsets.limit) {
michael@0:         return;
michael@0:     }
michael@0: 
michael@0:     // case type: >0 cased (UCASE_LOWER etc.)  ==0 uncased  <0 case-ignorable
michael@0:     int32_t type;
michael@0: 
michael@0:     // Our mode; we are either converting letter toTitle or
michael@0:     // toLower.
michael@0:     UBool doTitle = TRUE;
michael@0:     
michael@0:     // Determine if there is a preceding context of cased case-ignorable*,
michael@0:     // in which case we want to start in toLower mode.  If the
michael@0:     // prior context is anything else (including empty) then start
michael@0:     // in toTitle mode.
michael@0:     UChar32 c;
michael@0:     int32_t start;
michael@0:     for (start = offsets.start - 1; start >= offsets.contextStart; start -= U16_LENGTH(c)) {
michael@0:         c = text.char32At(start);
michael@0:         type=ucase_getTypeOrIgnorable(fCsp, c);
michael@0:         if(type>0) { // cased
michael@0:             doTitle=FALSE;
michael@0:             break;
michael@0:         } else if(type==0) { // uncased but not ignorable
michael@0:             break;
michael@0:         }
michael@0:         // else (type<0) case-ignorable: continue
michael@0:     }
michael@0:     
michael@0:     // Convert things after a cased character toLower; things
michael@0:     // after an uncased, non-case-ignorable character toTitle.  Case-ignorable
michael@0:     // characters are copied directly and do not change the mode.
michael@0:     UCaseContext csc;
michael@0:     uprv_memset(&csc, 0, sizeof(csc));
michael@0:     csc.p = &text;
michael@0:     csc.start = offsets.contextStart;
michael@0:     csc.limit = offsets.contextLimit;
michael@0: 
michael@0:     UnicodeString tmp;
michael@0:     const UChar *s;
michael@0:     int32_t textPos, delta, result, locCache=0;
michael@0: 
michael@0:     for(textPos=offsets.start; textPos<offsets.limit;) {
michael@0:         csc.cpStart=textPos;
michael@0:         c=text.char32At(textPos);
michael@0:         csc.cpLimit=textPos+=U16_LENGTH(c);
michael@0: 
michael@0:         type=ucase_getTypeOrIgnorable(fCsp, c);
michael@0:         if(type>=0) { // not case-ignorable
michael@0:             if(doTitle) {
michael@0:                 result=ucase_toFullTitle(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache);
michael@0:             } else {
michael@0:                 result=ucase_toFullLower(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache);
michael@0:             }
michael@0:             doTitle = (UBool)(type==0); // doTitle=isUncased
michael@0: 
michael@0:             if(csc.b1 && isIncremental) {
michael@0:                 // fMap() tried to look beyond the context limit
michael@0:                 // wait for more input
michael@0:                 offsets.start=csc.cpStart;
michael@0:                 return;
michael@0:             }
michael@0: 
michael@0:             if(result>=0) {
michael@0:                 // replace the current code point with its full case mapping result
michael@0:                 // see UCASE_MAX_STRING_LENGTH
michael@0:                 if(result<=UCASE_MAX_STRING_LENGTH) {
michael@0:                     // string s[result]
michael@0:                     tmp.setTo(FALSE, s, result);
michael@0:                     delta=result-U16_LENGTH(c);
michael@0:                 } else {
michael@0:                     // single code point
michael@0:                     tmp.setTo(result);
michael@0:                     delta=tmp.length()-U16_LENGTH(c);
michael@0:                 }
michael@0:                 text.handleReplaceBetween(csc.cpStart, textPos, tmp);
michael@0:                 if(delta!=0) {
michael@0:                     textPos+=delta;
michael@0:                     csc.limit=offsets.contextLimit+=delta;
michael@0:                     offsets.limit+=delta;
michael@0:                 }
michael@0:             }
michael@0:         }
michael@0:     }
michael@0:     offsets.start=textPos;
michael@0: }
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif /* #if !UCONFIG_NO_TRANSLITERATION */