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: *   11/19/2001  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/utf16.h"
michael@0: #include "esctrn.h"
michael@0: #include "util.h"
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: static const UChar UNIPRE[] = {85,43,0}; // "U+"
michael@0: static const UChar BS_u[] = {92,117,0}; // "\\u"
michael@0: static const UChar BS_U[] = {92,85,0}; // "\\U"
michael@0: static const UChar XMLPRE[] = {38,35,120,0}; // "&#x"
michael@0: static const UChar XML10PRE[] = {38,35,0}; // "&#"
michael@0: static const UChar PERLPRE[] = {92,120,123,0}; // "\\x{"
michael@0: static const UChar SEMI[] = {59,0}; // ";"
michael@0: static const UChar RBRACE[] = {125,0}; // "}"
michael@0: 
michael@0: UOBJECT_DEFINE_RTTI_IMPLEMENTATION(EscapeTransliterator)
michael@0: 
michael@0: /**
michael@0:  * Factory methods
michael@0:  */
michael@0: static Transliterator* _createEscUnicode(const UnicodeString& ID, Transliterator::Token /*context*/) {
michael@0:     // Unicode: "U+10FFFF" hex, min=4, max=6
michael@0:     return new EscapeTransliterator(ID, UnicodeString(TRUE, UNIPRE, 2), UnicodeString(), 16, 4, TRUE, NULL);
michael@0: }
michael@0: static Transliterator* _createEscJava(const UnicodeString& ID, Transliterator::Token /*context*/) {
michael@0:     // Java: "\\uFFFF" hex, min=4, max=4
michael@0:     return new EscapeTransliterator(ID, UnicodeString(TRUE, BS_u, 2), UnicodeString(), 16, 4, FALSE, NULL);
michael@0: }
michael@0: static Transliterator* _createEscC(const UnicodeString& ID, Transliterator::Token /*context*/) {
michael@0:     // C: "\\uFFFF" hex, min=4, max=4; \\U0010FFFF hex, min=8, max=8
michael@0:     return new EscapeTransliterator(ID, UnicodeString(TRUE, BS_u, 2), UnicodeString(), 16, 4, TRUE,
michael@0:              new EscapeTransliterator(UnicodeString(), UnicodeString(TRUE, BS_U, 2), UnicodeString(), 16, 8, TRUE, NULL));
michael@0: }
michael@0: static Transliterator* _createEscXML(const UnicodeString& ID, Transliterator::Token /*context*/) {
michael@0:     // XML: "" hex, min=1, max=6
michael@0:     return new EscapeTransliterator(ID, UnicodeString(TRUE, XMLPRE, 3), UnicodeString(SEMI[0]), 16, 1, TRUE, NULL);
michael@0: }
michael@0: static Transliterator* _createEscXML10(const UnicodeString& ID, Transliterator::Token /*context*/) {
michael@0:     // XML10: "&1114111;" dec, min=1, max=7 (not really "Any-Hex")
michael@0:     return new EscapeTransliterator(ID, UnicodeString(TRUE, XML10PRE, 2), UnicodeString(SEMI[0]), 10, 1, TRUE, NULL);
michael@0: }
michael@0: static Transliterator* _createEscPerl(const UnicodeString& ID, Transliterator::Token /*context*/) {
michael@0:     // Perl: "\\x{263A}" hex, min=1, max=6
michael@0:     return new EscapeTransliterator(ID, UnicodeString(TRUE, PERLPRE, 3), UnicodeString(RBRACE[0]), 16, 1, TRUE, NULL);
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Registers standard variants with the system.  Called by
michael@0:  * Transliterator during initialization.
michael@0:  */
michael@0: void EscapeTransliterator::registerIDs() {
michael@0:     Token t = integerToken(0);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex/Unicode"), _createEscUnicode, t);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex/Java"), _createEscJava, t);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex/C"), _createEscC, t);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex/XML"), _createEscXML, t);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex/XML10"), _createEscXML10, t);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex/Perl"), _createEscPerl, t);
michael@0: 
michael@0:     Transliterator::_registerFactory(UNICODE_STRING_SIMPLE("Any-Hex"), _createEscJava, t);
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Constructs an escape transliterator with the given ID and
michael@0:  * parameters.  See the class member documentation for details.
michael@0:  */
michael@0: EscapeTransliterator::EscapeTransliterator(const UnicodeString& newID,
michael@0:                          const UnicodeString& _prefix, const UnicodeString& _suffix,
michael@0:                          int32_t _radix, int32_t _minDigits,
michael@0:                          UBool _grokSupplementals,
michael@0:                          EscapeTransliterator* adoptedSupplementalHandler) :
michael@0:     Transliterator(newID, NULL)
michael@0: {
michael@0:     this->prefix = _prefix;
michael@0:     this->suffix = _suffix;
michael@0:     this->radix = _radix;
michael@0:     this->minDigits = _minDigits;
michael@0:     this->grokSupplementals = _grokSupplementals;
michael@0:     this->supplementalHandler = adoptedSupplementalHandler;
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Copy constructor.
michael@0:  */
michael@0: EscapeTransliterator::EscapeTransliterator(const EscapeTransliterator& o) :
michael@0:     Transliterator(o),
michael@0:     prefix(o.prefix),
michael@0:     suffix(o.suffix),
michael@0:     radix(o.radix),
michael@0:     minDigits(o.minDigits),
michael@0:     grokSupplementals(o.grokSupplementals) {
michael@0:     supplementalHandler = (o.supplementalHandler != 0) ?
michael@0:         new EscapeTransliterator(*o.supplementalHandler) : NULL;
michael@0: }
michael@0: 
michael@0: EscapeTransliterator::~EscapeTransliterator() {
michael@0:     delete supplementalHandler;
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Transliterator API.
michael@0:  */
michael@0: Transliterator* EscapeTransliterator::clone() const {
michael@0:     return new EscapeTransliterator(*this);
michael@0: }
michael@0: 
michael@0: /**
michael@0:  * Implements {@link Transliterator#handleTransliterate}.
michael@0:  */
michael@0: void EscapeTransliterator::handleTransliterate(Replaceable& text,
michael@0:                                                UTransPosition& pos,
michael@0:                                                UBool /*isIncremental*/) const
michael@0: {
michael@0:     /* TODO: Verify that isIncremental can be ignored */
michael@0:     int32_t start = pos.start;
michael@0:     int32_t limit = pos.limit;
michael@0: 
michael@0:     UnicodeString buf(prefix);
michael@0:     int32_t prefixLen = prefix.length();
michael@0:     UBool redoPrefix = FALSE;
michael@0: 
michael@0:     while (start < limit) {
michael@0:         int32_t c = grokSupplementals ? text.char32At(start) : text.charAt(start);
michael@0:         int32_t charLen = grokSupplementals ? U16_LENGTH(c) : 1;
michael@0: 
michael@0:         if ((c & 0xFFFF0000) != 0 && supplementalHandler != NULL) {
michael@0:             buf.truncate(0);
michael@0:             buf.append(supplementalHandler->prefix);
michael@0:             ICU_Utility::appendNumber(buf, c, supplementalHandler->radix,
michael@0:                                   supplementalHandler->minDigits);
michael@0:             buf.append(supplementalHandler->suffix);
michael@0:             redoPrefix = TRUE;
michael@0:         } else {
michael@0:             if (redoPrefix) {
michael@0:                 buf.truncate(0);
michael@0:                 buf.append(prefix);
michael@0:                 redoPrefix = FALSE;
michael@0:             } else {
michael@0:                 buf.truncate(prefixLen);
michael@0:             }
michael@0:             ICU_Utility::appendNumber(buf, c, radix, minDigits);
michael@0:             buf.append(suffix);
michael@0:         }
michael@0: 
michael@0:         text.handleReplaceBetween(start, start + charLen, buf);
michael@0:         start += buf.length();
michael@0:         limit += buf.length() - charLen;
michael@0:     }
michael@0: 
michael@0:     pos.contextLimit += limit - pos.limit;
michael@0:     pos.limit = limit;
michael@0:     pos.start = start;
michael@0: }
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif /* #if !UCONFIG_NO_TRANSLITERATION */
michael@0: 
michael@0: //eof