michael@0: /*
michael@0: *******************************************************************************
michael@0: * Copyright (C) 2009-2013, International Business Machines Corporation and
michael@0: * others. All Rights Reserved.
michael@0: *******************************************************************************
michael@0: *
michael@0: * File PLURFMT.CPP
michael@0: *******************************************************************************
michael@0: */
michael@0: 
michael@0: #include "unicode/decimfmt.h"
michael@0: #include "unicode/messagepattern.h"
michael@0: #include "unicode/plurfmt.h"
michael@0: #include "unicode/plurrule.h"
michael@0: #include "unicode/utypes.h"
michael@0: #include "cmemory.h"
michael@0: #include "messageimpl.h"
michael@0: #include "plurrule_impl.h"
michael@0: #include "uassert.h"
michael@0: #include "uhash.h"
michael@0: 
michael@0: #if !UCONFIG_NO_FORMATTING
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: static const UChar OTHER_STRING[] = {
michael@0:     0x6F, 0x74, 0x68, 0x65, 0x72, 0  // "other"
michael@0: };
michael@0: 
michael@0: UOBJECT_DEFINE_RTTI_IMPLEMENTATION(PluralFormat)
michael@0: 
michael@0: PluralFormat::PluralFormat(UErrorCode& status)
michael@0:         : locale(Locale::getDefault()),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(NULL, UPLURAL_TYPE_CARDINAL, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const Locale& loc, UErrorCode& status)
michael@0:         : locale(loc),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(NULL, UPLURAL_TYPE_CARDINAL, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const PluralRules& rules, UErrorCode& status)
michael@0:         : locale(Locale::getDefault()),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(&rules, UPLURAL_TYPE_COUNT, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const Locale& loc,
michael@0:                            const PluralRules& rules,
michael@0:                            UErrorCode& status)
michael@0:         : locale(loc),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(&rules, UPLURAL_TYPE_COUNT, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const Locale& loc,
michael@0:                            UPluralType type,
michael@0:                            UErrorCode& status)
michael@0:         : locale(loc),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(NULL, type, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const UnicodeString& pat,
michael@0:                            UErrorCode& status)
michael@0:         : locale(Locale::getDefault()),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(NULL, UPLURAL_TYPE_CARDINAL, status);
michael@0:     applyPattern(pat, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const Locale& loc,
michael@0:                            const UnicodeString& pat,
michael@0:                            UErrorCode& status)
michael@0:         : locale(loc),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(NULL, UPLURAL_TYPE_CARDINAL, status);
michael@0:     applyPattern(pat, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const PluralRules& rules,
michael@0:                            const UnicodeString& pat,
michael@0:                            UErrorCode& status)
michael@0:         : locale(Locale::getDefault()),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(&rules, UPLURAL_TYPE_COUNT, status);
michael@0:     applyPattern(pat, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const Locale& loc,
michael@0:                            const PluralRules& rules,
michael@0:                            const UnicodeString& pat,
michael@0:                            UErrorCode& status)
michael@0:         : locale(loc),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(&rules, UPLURAL_TYPE_COUNT, status);
michael@0:     applyPattern(pat, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const Locale& loc,
michael@0:                            UPluralType type,
michael@0:                            const UnicodeString& pat,
michael@0:                            UErrorCode& status)
michael@0:         : locale(loc),
michael@0:           msgPattern(status),
michael@0:           numberFormat(NULL),
michael@0:           offset(0) {
michael@0:     init(NULL, type, status);
michael@0:     applyPattern(pat, status);
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralFormat(const PluralFormat& other)
michael@0:         : Format(other),
michael@0:           locale(other.locale),
michael@0:           msgPattern(other.msgPattern),
michael@0:           numberFormat(NULL),
michael@0:           offset(other.offset) {
michael@0:     copyObjects(other);
michael@0: }
michael@0: 
michael@0: void
michael@0: PluralFormat::copyObjects(const PluralFormat& other) {
michael@0:     UErrorCode status = U_ZERO_ERROR;
michael@0:     if (numberFormat != NULL) {
michael@0:         delete numberFormat;
michael@0:     }
michael@0:     if (pluralRulesWrapper.pluralRules != NULL) {
michael@0:         delete pluralRulesWrapper.pluralRules;
michael@0:     }
michael@0: 
michael@0:     if (other.numberFormat == NULL) {
michael@0:         numberFormat = NumberFormat::createInstance(locale, status);
michael@0:     } else {
michael@0:         numberFormat = (NumberFormat*)other.numberFormat->clone();
michael@0:     }
michael@0:     if (other.pluralRulesWrapper.pluralRules == NULL) {
michael@0:         pluralRulesWrapper.pluralRules = PluralRules::forLocale(locale, status);
michael@0:     } else {
michael@0:         pluralRulesWrapper.pluralRules = other.pluralRulesWrapper.pluralRules->clone();
michael@0:     }
michael@0: }
michael@0: 
michael@0: 
michael@0: PluralFormat::~PluralFormat() {
michael@0:     delete numberFormat;
michael@0: }
michael@0: 
michael@0: void
michael@0: PluralFormat::init(const PluralRules* rules, UPluralType type, UErrorCode& status) {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return;
michael@0:     }
michael@0: 
michael@0:     if (rules==NULL) {
michael@0:         pluralRulesWrapper.pluralRules = PluralRules::forLocale(locale, type, status);
michael@0:     } else {
michael@0:         pluralRulesWrapper.pluralRules = rules->clone();
michael@0:         if (pluralRulesWrapper.pluralRules == NULL) {
michael@0:             status = U_MEMORY_ALLOCATION_ERROR;
michael@0:             return;
michael@0:         }
michael@0:     }
michael@0: 
michael@0:     numberFormat= NumberFormat::createInstance(locale, status);
michael@0: }
michael@0: 
michael@0: void
michael@0: PluralFormat::applyPattern(const UnicodeString& newPattern, UErrorCode& status) {
michael@0:     msgPattern.parsePluralStyle(newPattern, NULL, status);
michael@0:     if (U_FAILURE(status)) {
michael@0:         msgPattern.clear();
michael@0:         offset = 0;
michael@0:         return;
michael@0:     }
michael@0:     offset = msgPattern.getPluralOffset(0);
michael@0: }
michael@0: 
michael@0: UnicodeString&
michael@0: PluralFormat::format(const Formattable& obj,
michael@0:                    UnicodeString& appendTo,
michael@0:                    FieldPosition& pos,
michael@0:                    UErrorCode& status) const
michael@0: {
michael@0:     if (U_FAILURE(status)) return appendTo;
michael@0: 
michael@0:     if (obj.isNumeric()) {
michael@0:         return format(obj, obj.getDouble(), appendTo, pos, status);
michael@0:     } else {
michael@0:         status = U_ILLEGAL_ARGUMENT_ERROR;
michael@0:         return appendTo;
michael@0:     }
michael@0: }
michael@0: 
michael@0: UnicodeString
michael@0: PluralFormat::format(int32_t number, UErrorCode& status) const {
michael@0:     FieldPosition fpos(0);
michael@0:     UnicodeString result;
michael@0:     return format(Formattable(number), number, result, fpos, status);
michael@0: }
michael@0: 
michael@0: UnicodeString
michael@0: PluralFormat::format(double number, UErrorCode& status) const {
michael@0:     FieldPosition fpos(0);
michael@0:     UnicodeString result;
michael@0:     return format(Formattable(number), number, result, fpos, status);
michael@0: }
michael@0: 
michael@0: 
michael@0: UnicodeString&
michael@0: PluralFormat::format(int32_t number,
michael@0:                      UnicodeString& appendTo,
michael@0:                      FieldPosition& pos,
michael@0:                      UErrorCode& status) const {
michael@0:     return format(Formattable(number), (double)number, appendTo, pos, status);
michael@0: }
michael@0: 
michael@0: UnicodeString&
michael@0: PluralFormat::format(double number,
michael@0:                      UnicodeString& appendTo,
michael@0:                      FieldPosition& pos,
michael@0:                      UErrorCode& status) const {
michael@0:     return format(Formattable(number), (double)number, appendTo, pos, status);
michael@0: }
michael@0: 
michael@0: UnicodeString&
michael@0: PluralFormat::format(const Formattable& numberObject, double number,
michael@0:                      UnicodeString& appendTo,
michael@0:                      FieldPosition& pos,
michael@0:                      UErrorCode& status) const {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return appendTo;
michael@0:     }
michael@0:     if (msgPattern.countParts() == 0) {
michael@0:         return numberFormat->format(numberObject, appendTo, pos, status);
michael@0:     }
michael@0:     // Get the appropriate sub-message.
michael@0:     // Select it based on the formatted number-offset.
michael@0:     double numberMinusOffset = number - offset;
michael@0:     UnicodeString numberString;
michael@0:     FieldPosition ignorePos;
michael@0:     FixedDecimal dec(numberMinusOffset);
michael@0:     if (offset == 0) {
michael@0:         numberFormat->format(numberObject, numberString, ignorePos, status);  // could be BigDecimal etc.
michael@0:         DecimalFormat *decFmt = dynamic_cast<DecimalFormat *>(numberFormat);
michael@0:         if(decFmt != NULL) {
michael@0:             dec = decFmt->getFixedDecimal(numberObject, status);
michael@0:         }
michael@0:     } else {
michael@0:         numberFormat->format(numberMinusOffset, numberString, ignorePos, status);
michael@0:         DecimalFormat *decFmt = dynamic_cast<DecimalFormat *>(numberFormat);
michael@0:         if(decFmt != NULL) {
michael@0:             dec = decFmt->getFixedDecimal(numberMinusOffset, status);
michael@0:         }
michael@0:     }
michael@0:     int32_t partIndex = findSubMessage(msgPattern, 0, pluralRulesWrapper, &dec, number, status);
michael@0:     if (U_FAILURE(status)) { return appendTo; }
michael@0:     // Replace syntactic # signs in the top level of this sub-message
michael@0:     // (not in nested arguments) with the formatted number-offset.
michael@0:     const UnicodeString& pattern = msgPattern.getPatternString();
michael@0:     int32_t prevIndex = msgPattern.getPart(partIndex).getLimit();
michael@0:     for (;;) {
michael@0:         const MessagePattern::Part& part = msgPattern.getPart(++partIndex);
michael@0:         const UMessagePatternPartType type = part.getType();
michael@0:         int32_t index = part.getIndex();
michael@0:         if (type == UMSGPAT_PART_TYPE_MSG_LIMIT) {
michael@0:             return appendTo.append(pattern, prevIndex, index - prevIndex);
michael@0:         } else if ((type == UMSGPAT_PART_TYPE_REPLACE_NUMBER) ||
michael@0:             (type == UMSGPAT_PART_TYPE_SKIP_SYNTAX && MessageImpl::jdkAposMode(msgPattern))) {
michael@0:             appendTo.append(pattern, prevIndex, index - prevIndex);
michael@0:             if (type == UMSGPAT_PART_TYPE_REPLACE_NUMBER) {
michael@0:                 appendTo.append(numberString);
michael@0:             }
michael@0:             prevIndex = part.getLimit();
michael@0:         } else if (type == UMSGPAT_PART_TYPE_ARG_START) {
michael@0:             appendTo.append(pattern, prevIndex, index - prevIndex);
michael@0:             prevIndex = index;
michael@0:             partIndex = msgPattern.getLimitPartIndex(partIndex);
michael@0:             index = msgPattern.getPart(partIndex).getLimit();
michael@0:             MessageImpl::appendReducedApostrophes(pattern, prevIndex, index, appendTo);
michael@0:             prevIndex = index;
michael@0:         }
michael@0:     }
michael@0: }
michael@0: 
michael@0: UnicodeString&
michael@0: PluralFormat::toPattern(UnicodeString& appendTo) {
michael@0:     if (0 == msgPattern.countParts()) {
michael@0:         appendTo.setToBogus();
michael@0:     } else {
michael@0:         appendTo.append(msgPattern.getPatternString());
michael@0:     }
michael@0:     return appendTo;
michael@0: }
michael@0: 
michael@0: void
michael@0: PluralFormat::setLocale(const Locale& loc, UErrorCode& status) {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return;
michael@0:     }
michael@0:     locale = loc;
michael@0:     msgPattern.clear();
michael@0:     delete numberFormat;
michael@0:     offset = 0;
michael@0:     numberFormat = NULL;
michael@0:     pluralRulesWrapper.reset();
michael@0:     init(NULL, UPLURAL_TYPE_CARDINAL, status);
michael@0: }
michael@0: 
michael@0: void
michael@0: PluralFormat::setNumberFormat(const NumberFormat* format, UErrorCode& status) {
michael@0:     if (U_FAILURE(status)) {
michael@0:         return;
michael@0:     }
michael@0:     NumberFormat* nf = (NumberFormat*)format->clone();
michael@0:     if (nf != NULL) {
michael@0:         delete numberFormat;
michael@0:         numberFormat = nf;
michael@0:     } else {
michael@0:         status = U_MEMORY_ALLOCATION_ERROR;
michael@0:     }
michael@0: }
michael@0: 
michael@0: Format*
michael@0: PluralFormat::clone() const
michael@0: {
michael@0:     return new PluralFormat(*this);
michael@0: }
michael@0: 
michael@0: 
michael@0: PluralFormat&
michael@0: PluralFormat::operator=(const PluralFormat& other) {
michael@0:     if (this != &other) {
michael@0:         locale = other.locale;
michael@0:         msgPattern = other.msgPattern;
michael@0:         offset = other.offset;
michael@0:         copyObjects(other);
michael@0:     }
michael@0: 
michael@0:     return *this;
michael@0: }
michael@0: 
michael@0: UBool
michael@0: PluralFormat::operator==(const Format& other) const {
michael@0:     if (this == &other) {
michael@0:         return TRUE;
michael@0:     }
michael@0:     if (!Format::operator==(other)) {
michael@0:         return FALSE;
michael@0:     }
michael@0:     const PluralFormat& o = (const PluralFormat&)other;
michael@0:     return
michael@0:         locale == o.locale &&
michael@0:         msgPattern == o.msgPattern &&  // implies same offset
michael@0:         (numberFormat == NULL) == (o.numberFormat == NULL) &&
michael@0:         (numberFormat == NULL || *numberFormat == *o.numberFormat) &&
michael@0:         (pluralRulesWrapper.pluralRules == NULL) == (o.pluralRulesWrapper.pluralRules == NULL) &&
michael@0:         (pluralRulesWrapper.pluralRules == NULL ||
michael@0:             *pluralRulesWrapper.pluralRules == *o.pluralRulesWrapper.pluralRules);
michael@0: }
michael@0: 
michael@0: UBool
michael@0: PluralFormat::operator!=(const Format& other) const {
michael@0:     return  !operator==(other);
michael@0: }
michael@0: 
michael@0: void
michael@0: PluralFormat::parseObject(const UnicodeString& /*source*/,
michael@0:                         Formattable& /*result*/,
michael@0:                         ParsePosition& pos) const
michael@0: {
michael@0:     // Parsing not supported.
michael@0:     pos.setErrorIndex(pos.getIndex());
michael@0: }
michael@0: 
michael@0: int32_t PluralFormat::findSubMessage(const MessagePattern& pattern, int32_t partIndex,
michael@0:                                      const PluralSelector& selector, void *context,
michael@0:                                      double number, UErrorCode& ec) {
michael@0:     if (U_FAILURE(ec)) {
michael@0:         return 0;
michael@0:     }
michael@0:     int32_t count=pattern.countParts();
michael@0:     double offset;
michael@0:     const MessagePattern::Part* part=&pattern.getPart(partIndex);
michael@0:     if (MessagePattern::Part::hasNumericValue(part->getType())) {
michael@0:         offset=pattern.getNumericValue(*part);
michael@0:         ++partIndex;
michael@0:     } else {
michael@0:         offset=0;
michael@0:     }
michael@0:     // The keyword is empty until we need to match against a non-explicit, not-"other" value.
michael@0:     // Then we get the keyword from the selector.
michael@0:     // (In other words, we never call the selector if we match against an explicit value,
michael@0:     // or if the only non-explicit keyword is "other".)
michael@0:     UnicodeString keyword;
michael@0:     UnicodeString other(FALSE, OTHER_STRING, 5);
michael@0:     // When we find a match, we set msgStart>0 and also set this boolean to true
michael@0:     // to avoid matching the keyword again (duplicates are allowed)
michael@0:     // while we continue to look for an explicit-value match.
michael@0:     UBool haveKeywordMatch=FALSE;
michael@0:     // msgStart is 0 until we find any appropriate sub-message.
michael@0:     // We remember the first "other" sub-message if we have not seen any
michael@0:     // appropriate sub-message before.
michael@0:     // We remember the first matching-keyword sub-message if we have not seen
michael@0:     // one of those before.
michael@0:     // (The parser allows [does not check for] duplicate keywords.
michael@0:     // We just have to make sure to take the first one.)
michael@0:     // We avoid matching the keyword twice by also setting haveKeywordMatch=true
michael@0:     // at the first keyword match.
michael@0:     // We keep going until we find an explicit-value match or reach the end of the plural style.
michael@0:     int32_t msgStart=0;
michael@0:     // Iterate over (ARG_SELECTOR [ARG_INT|ARG_DOUBLE] message) tuples
michael@0:     // until ARG_LIMIT or end of plural-only pattern.
michael@0:     do {
michael@0:         part=&pattern.getPart(partIndex++);
michael@0:         const UMessagePatternPartType type = part->getType();
michael@0:         if(type==UMSGPAT_PART_TYPE_ARG_LIMIT) {
michael@0:             break;
michael@0:         }
michael@0:         U_ASSERT (type==UMSGPAT_PART_TYPE_ARG_SELECTOR);
michael@0:         // part is an ARG_SELECTOR followed by an optional explicit value, and then a message
michael@0:         if(MessagePattern::Part::hasNumericValue(pattern.getPartType(partIndex))) {
michael@0:             // explicit value like "=2"
michael@0:             part=&pattern.getPart(partIndex++);
michael@0:             if(number==pattern.getNumericValue(*part)) {
michael@0:                 // matches explicit value
michael@0:                 return partIndex;
michael@0:             }
michael@0:         } else if(!haveKeywordMatch) {
michael@0:             // plural keyword like "few" or "other"
michael@0:             // Compare "other" first and call the selector if this is not "other".
michael@0:             if(pattern.partSubstringMatches(*part, other)) {
michael@0:                 if(msgStart==0) {
michael@0:                     msgStart=partIndex;
michael@0:                     if(0 == keyword.compare(other)) {
michael@0:                         // This is the first "other" sub-message,
michael@0:                         // and the selected keyword is also "other".
michael@0:                         // Do not match "other" again.
michael@0:                         haveKeywordMatch=TRUE;
michael@0:                     }
michael@0:                 }
michael@0:             } else {
michael@0:                 if(keyword.isEmpty()) {
michael@0:                     keyword=selector.select(context, number-offset, ec);
michael@0:                     if(msgStart!=0 && (0 == keyword.compare(other))) {
michael@0:                         // We have already seen an "other" sub-message.
michael@0:                         // Do not match "other" again.
michael@0:                         haveKeywordMatch=TRUE;
michael@0:                         // Skip keyword matching but do getLimitPartIndex().
michael@0:                     }
michael@0:                 }
michael@0:                 if(!haveKeywordMatch && pattern.partSubstringMatches(*part, keyword)) {
michael@0:                     // keyword matches
michael@0:                     msgStart=partIndex;
michael@0:                     // Do not match this keyword again.
michael@0:                     haveKeywordMatch=TRUE;
michael@0:                 }
michael@0:             }
michael@0:         }
michael@0:         partIndex=pattern.getLimitPartIndex(partIndex);
michael@0:     } while(++partIndex<count);
michael@0:     return msgStart;
michael@0: }
michael@0: 
michael@0: PluralFormat::PluralSelector::~PluralSelector() {}
michael@0: 
michael@0: PluralFormat::PluralSelectorAdapter::~PluralSelectorAdapter() {
michael@0:     delete pluralRules;
michael@0: }
michael@0: 
michael@0: UnicodeString PluralFormat::PluralSelectorAdapter::select(void *context, double number,
michael@0:                                                           UErrorCode& /*ec*/) const {
michael@0:     (void)number;  // unused except in the assertion
michael@0:     FixedDecimal *dec=static_cast<FixedDecimal *>(context);
michael@0:     U_ASSERT(dec->source==number);
michael@0:     return pluralRules->select(*dec);
michael@0: }
michael@0: 
michael@0: void PluralFormat::PluralSelectorAdapter::reset() {
michael@0:     delete pluralRules;
michael@0:     pluralRules = NULL;
michael@0: }
michael@0: 
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: 
michael@0: #endif /* #if !UCONFIG_NO_FORMATTING */
michael@0: 
michael@0: //eof