/** * Get a set containing the expansions defined by the collator. The set includes * both the UCA expansions and the expansions defined by the tailoring * @param coll collator * @param conts the set to hold the result * @param addPrefixes add the prefix contextual elements to contractions * @param status to hold the error code * * @draft ICU 3.4 */ U_CAPI void U_EXPORT2 ucol_getContractionsAndExpansions( const UCollator *coll, USet *contractions, USet *expansions, UBool addPrefixes, UErrorCode *status) { if(U_FAILURE(*status)) { return; } if(coll == NULL) { *status = U_ILLEGAL_ARGUMENT_ERROR; return; } if(contractions) { uset_clear(contractions); } if(expansions) { uset_clear(expansions); } int32_t rulesLen = 0; const UChar* rules = ucol_getRules(coll, &rulesLen); UColTokenParser src; ucol_tok_initTokenList(&src, rules, rulesLen, coll->UCA, ucol_tok_getRulesFromBundle, NULL, status); contContext c = { NULL, contractions, expansions, src.removeSet, addPrefixes, status }; // Add the UCA contractions c.coll = coll->UCA; utrie_enum(&coll->UCA->mapping, NULL, _processSpecials, &c); // This is collator specific. Add contractions from a collator c.coll = coll; c.removedContractions = NULL; utrie_enum(&coll->mapping, NULL, _processSpecials, &c); ucol_tok_closeTokenList(&src); }
U_CAPI UBool U_EXPORT2 ucol_equals(const UCollator *source, const UCollator *target) { UErrorCode status = U_ZERO_ERROR; // if pointers are equal, collators are equal if(source == target) { return TRUE; } int32_t i = 0, j = 0; // if any of attributes are different, collators are not equal for(i = 0; i < UCOL_ATTRIBUTE_COUNT; i++) { if(ucol_getAttribute(source, (UColAttribute)i, &status) != ucol_getAttribute(target, (UColAttribute)i, &status) || U_FAILURE(status)) { return FALSE; } } int32_t sourceRulesLen = 0, targetRulesLen = 0; const UChar *sourceRules = ucol_getRules(source, &sourceRulesLen); const UChar *targetRules = ucol_getRules(target, &targetRulesLen); if(sourceRulesLen == targetRulesLen && u_strncmp(sourceRules, targetRules, sourceRulesLen) == 0) { // all the attributes are equal and the rules are equal - collators are equal return(TRUE); } // hard part, need to construct tree from rules and see if they yield the same tailoring UBool result = TRUE; UParseError parseError; UColTokenParser sourceParser, targetParser; int32_t sourceListLen = 0, targetListLen = 0; ucol_tok_initTokenList(&sourceParser, sourceRules, sourceRulesLen, source->UCA, &status); ucol_tok_initTokenList(&targetParser, targetRules, targetRulesLen, target->UCA, &status); sourceListLen = ucol_tok_assembleTokenList(&sourceParser, &parseError, &status); targetListLen = ucol_tok_assembleTokenList(&targetParser, &parseError, &status); if(sourceListLen != targetListLen) { // different number of resets result = FALSE; } else { UColToken *sourceReset = NULL, *targetReset = NULL; UChar *sourceResetString = NULL, *targetResetString = NULL; int32_t sourceStringLen = 0, targetStringLen = 0; for(i = 0; i < sourceListLen; i++) { sourceReset = sourceParser.lh[i].reset; sourceResetString = sourceParser.source+(sourceReset->source & 0xFFFFFF); sourceStringLen = sourceReset->source >> 24; for(j = 0; j < sourceListLen; j++) { targetReset = targetParser.lh[j].reset; targetResetString = targetParser.source+(targetReset->source & 0xFFFFFF); targetStringLen = targetReset->source >> 24; if(sourceStringLen == targetStringLen && (u_strncmp(sourceResetString, targetResetString, sourceStringLen) == 0)) { sourceReset = sourceParser.lh[i].first; targetReset = targetParser.lh[j].first; while(sourceReset != NULL && targetReset != NULL) { sourceResetString = sourceParser.source+(sourceReset->source & 0xFFFFFF); sourceStringLen = sourceReset->source >> 24; targetResetString = targetParser.source+(targetReset->source & 0xFFFFFF); targetStringLen = targetReset->source >> 24; if(sourceStringLen != targetStringLen || (u_strncmp(sourceResetString, targetResetString, sourceStringLen) != 0)) { result = FALSE; goto returnResult; } // probably also need to check the expansions if(sourceReset->expansion) { if(!targetReset->expansion) { result = FALSE; goto returnResult; } else { // compare expansions sourceResetString = sourceParser.source+(sourceReset->expansion& 0xFFFFFF); sourceStringLen = sourceReset->expansion >> 24; targetResetString = targetParser.source+(targetReset->expansion & 0xFFFFFF); targetStringLen = targetReset->expansion >> 24; if(sourceStringLen != targetStringLen || (u_strncmp(sourceResetString, targetResetString, sourceStringLen) != 0)) { result = FALSE; goto returnResult; } } } else { if(targetReset->expansion) { result = FALSE; goto returnResult; } } sourceReset = sourceReset->next; targetReset = targetReset->next; } if(sourceReset != targetReset) { // at least one is not NULL // there are more tailored elements in one list result = FALSE; goto returnResult; } break; } } // couldn't find the reset anchor, so the collators are not equal if(j == sourceListLen) { result = FALSE; goto returnResult; } }
U_CAPI UCollator* U_EXPORT2 ucol_openRules( const UChar *rules, int32_t rulesLength, UColAttributeValue normalizationMode, UCollationStrength strength, UParseError *parseError, UErrorCode *status) { UColTokenParser src; UColAttributeValue norm; UParseError tErr; if(status == NULL || U_FAILURE(*status)){ return 0; } u_init(status); if (U_FAILURE(*status)) { return NULL; } if(rules == NULL || rulesLength < -1) { *status = U_ILLEGAL_ARGUMENT_ERROR; return 0; } if(rulesLength == -1) { rulesLength = u_strlen(rules); } if(parseError == NULL){ parseError = &tErr; } switch(normalizationMode) { case UCOL_OFF: case UCOL_ON: case UCOL_DEFAULT: norm = normalizationMode; break; default: *status = U_ILLEGAL_ARGUMENT_ERROR; return 0; } UCollator *UCA = ucol_initUCA(status); if(U_FAILURE(*status)){ return NULL; } ucol_tok_initTokenList(&src, rules, rulesLength, UCA, status); ucol_tok_assembleTokenList(&src,parseError, status); if(U_FAILURE(*status)) { /* if status is U_ILLEGAL_ARGUMENT_ERROR, src->current points at the offending option */ /* if status is U_INVALID_FORMAT_ERROR, src->current points after the problematic part of the rules */ /* so something might be done here... or on lower level */ #ifdef UCOL_DEBUG if(*status == U_ILLEGAL_ARGUMENT_ERROR) { fprintf(stderr, "bad option starting at offset %i\n", src.current-src.source); } else { fprintf(stderr, "invalid rule just before offset %i\n", src.current-src.source); } #endif ucol_tok_closeTokenList(&src); return NULL; } UCollator *result = NULL; UCATableHeader *table = NULL; if(src.resultLen > 0 || src.removeSet != NULL) { /* we have a set of rules, let's make something of it */ /* also, if we wanted to remove some contractions, we should make a tailoring */ table = ucol_assembleTailoringTable(&src, status); if(U_SUCCESS(*status)) { // builder version table->version[0] = UCOL_BUILDER_VERSION; // no tailoring information on this level table->version[1] = table->version[2] = table->version[3] = 0; // set UCD version u_getUnicodeVersion(table->UCDVersion); // set UCA version uprv_memcpy(table->UCAVersion, UCA->image->UCAVersion, sizeof(UVersionInfo)); result = ucol_initCollator(table, 0, UCA, status); result->hasRealData = TRUE; result->freeImageOnClose = TRUE; } } else { /* no rules, but no error either */ // must be only options // We will init the collator from UCA result = ucol_initCollator(UCA->image, 0, UCA, status); // And set only the options UColOptionSet *opts = (UColOptionSet *)uprv_malloc(sizeof(UColOptionSet)); /* test for NULL */ if (opts == NULL) { *status = U_MEMORY_ALLOCATION_ERROR; goto cleanup; } uprv_memcpy(opts, src.opts, sizeof(UColOptionSet)); ucol_setOptionsFromHeader(result, opts, status); result->freeOptionsOnClose = TRUE; result->hasRealData = FALSE; result->freeImageOnClose = FALSE; } if(U_SUCCESS(*status)) { UChar *newRules; result->dataVersion[0] = UCOL_BUILDER_VERSION; if(rulesLength > 0) { newRules = (UChar *)uprv_malloc((rulesLength+1)*U_SIZEOF_UCHAR); /* test for NULL */ if (newRules == NULL) { *status = U_MEMORY_ALLOCATION_ERROR; goto cleanup; } uprv_memcpy(newRules, rules, rulesLength*U_SIZEOF_UCHAR); newRules[rulesLength]=0; result->rules = newRules; result->rulesLength = rulesLength; result->freeRulesOnClose = TRUE; } result->rb = NULL; result->elements = NULL; result->validLocale = NULL; result->requestedLocale = NULL; ucol_setAttribute(result, UCOL_STRENGTH, strength, status); ucol_setAttribute(result, UCOL_NORMALIZATION_MODE, norm, status); } else { cleanup: if(result != NULL) { ucol_close(result); } else { if(table != NULL) { uprv_free(table); } } result = NULL; } ucol_tok_closeTokenList(&src); return result; }