michael@0: /*
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: *******************************************************************************
michael@0: *   file name:  casetrn.cpp
michael@0: *   encoding:   US-ASCII
michael@0: *   tab size:   8 (not used)
michael@0: *   indentation:4
michael@0: *
michael@0: *   created on: 2004sep03
michael@0: *   created by: Markus W. Scherer
michael@0: *
michael@0: *   Implementation class for lower-/upper-/title-casing transliterators.
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/ustring.h"
michael@0: #include "unicode/utf.h"
michael@0: #include "unicode/utf16.h"
michael@0: #include "tolowtrn.h"
michael@0: #include "ucase.h"
michael@0: #include "cpputils.h"
michael@0: 
michael@0: /* case context iterator using a Replaceable */
michael@0: U_CFUNC UChar32 U_CALLCONV
michael@0: utrans_rep_caseContextIterator(void *context, int8_t dir)
michael@0: {
michael@0:     U_NAMESPACE_USE
michael@0: 
michael@0:     UCaseContext *csc=(UCaseContext *)context;
michael@0:     Replaceable *rep=(Replaceable *)csc->p;
michael@0:     UChar32 c;
michael@0: 
michael@0:     if(dir<0) {
michael@0:         /* reset for backward iteration */
michael@0:         csc->index=csc->cpStart;
michael@0:         csc->dir=dir;
michael@0:     } else if(dir>0) {
michael@0:         /* reset for forward iteration */
michael@0:         csc->index=csc->cpLimit;
michael@0:         csc->dir=dir;
michael@0:     } else {
michael@0:         /* continue current iteration direction */
michael@0:         dir=csc->dir;
michael@0:     }
michael@0: 
michael@0:     // automatically adjust start and limit if the Replaceable disagrees
michael@0:     // with the original values
michael@0:     if(dir<0) {
michael@0:         if(csc->start<csc->index) {
michael@0:             c=rep->char32At(csc->index-1);
michael@0:             if(c<0) {
michael@0:                 csc->start=csc->index;
michael@0:             } else {
michael@0:                 csc->index-=U16_LENGTH(c);
michael@0:                 return c;
michael@0:             }
michael@0:         }
michael@0:     } else {
michael@0:         // detect, and store in csc->b1, if we hit the limit
michael@0:         if(csc->index<csc->limit) {
michael@0:             c=rep->char32At(csc->index);
michael@0:             if(c<0) {
michael@0:                 csc->limit=csc->index;
michael@0:                 csc->b1=TRUE;
michael@0:             } else {
michael@0:                 csc->index+=U16_LENGTH(c);
michael@0:                 return c;
michael@0:             }
michael@0:         } else {
michael@0:             csc->b1=TRUE;
michael@0:         }
michael@0:     }
michael@0:     return U_SENTINEL;
michael@0: }
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: UOBJECT_DEFINE_ABSTRACT_RTTI_IMPLEMENTATION(CaseMapTransliterator)
michael@0: 
michael@0: /**
michael@0:  * Constructs a transliterator.
michael@0:  */
michael@0: CaseMapTransliterator::CaseMapTransliterator(const UnicodeString &id, UCaseMapFull *map) : 
michael@0:     Transliterator(id, 0),
michael@0:     fCsp(ucase_getSingleton()),
michael@0:     fMap(map)
michael@0: {
michael@0:     // TODO test incremental mode with context-sensitive text (e.g. greek sigma)
michael@0:     // TODO need to call setMaximumContextLength()?!
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Destructor.
michael@0:  */
michael@0: CaseMapTransliterator::~CaseMapTransliterator() {
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Copy constructor.
michael@0:  */
michael@0: CaseMapTransliterator::CaseMapTransliterator(const CaseMapTransliterator& o) :
michael@0:     Transliterator(o),
michael@0:     fCsp(o.fCsp), fMap(o.fMap)
michael@0: {
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Assignment operator.
michael@0:  */
michael@0: /*CaseMapTransliterator& CaseMapTransliterator::operator=(const CaseMapTransliterator& o) {
michael@0:     Transliterator::operator=(o);
michael@0:     fCsp = o.fCsp;
michael@0:     fMap = o.fMap;
michael@0:     return *this;
michael@0: }*/
michael@0: 
michael@0: /**
michael@0:  * Transliterator API.
michael@0:  */
michael@0: /*Transliterator* CaseMapTransliterator::clone(void) const {
michael@0:     return new CaseMapTransliterator(*this);
michael@0: }*/
michael@0: 
michael@0: /**
michael@0:  * Implements {@link Transliterator#handleTransliterate}.
michael@0:  */
michael@0: void CaseMapTransliterator::handleTransliterate(Replaceable& text,
michael@0:                                  UTransPosition& offsets, 
michael@0:                                  UBool isIncremental) const
michael@0: {
michael@0:     if (offsets.start >= offsets.limit) {
michael@0:         return;
michael@0:     }
michael@0: 
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:     UChar32 c;
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:         result=fMap(fCsp, c, utrans_rep_caseContextIterator, &csc, &s, "", &locCache);
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:     offsets.start=textPos;
michael@0: }
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif /* #if !UCONFIG_NO_TRANSLITERATION */