2 *******************************************************************************
4 * Copyright (C) 2011, International Business Machines
5 * Corporation and others. All Rights Reserved.
7 *******************************************************************************
8 * file name: uniset_closure.cpp
10 * tab size: 8 (not used)
13 * created on: 2011may30
14 * created by: Markus W. Scherer
16 * UnicodeSet::closeOver() and related methods moved here from uniset_props.cpp
17 * to simplify dependencies.
18 * In particular, this depends on the BreakIterator, but the BreakIterator
19 * code also builds UnicodeSets from patterns and needs uniset_props.
22 #include "unicode/brkiter.h"
23 #include "unicode/locid.h"
24 #include "unicode/parsepos.h"
25 #include "unicode/uniset.h"
32 // initial storage. Must be >= 0
33 // *** same as in uniset.cpp ! ***
34 #define START_EXTRA 16
38 // TODO memory debugging provided inside uniset.cpp
39 // could be made available here but probably obsolete with use of modern
40 // memory leak checker tools
43 //----------------------------------------------------------------
45 //----------------------------------------------------------------
47 UnicodeSet::UnicodeSet(const UnicodeString& pattern,
49 const SymbolTable* symbols,
51 len(0), capacity(START_EXTRA), list(0), bmpSet(0), buffer(0),
52 bufferCapacity(0), patLen(0), pat(NULL), strings(NULL), stringSpan(NULL),
55 if(U_SUCCESS(status)){
56 list = (UChar32*) uprv_malloc(sizeof(UChar32) * capacity);
59 status = U_MEMORY_ALLOCATION_ERROR;
61 allocateStrings(status);
62 applyPattern(pattern, options, symbols, status);
68 UnicodeSet::UnicodeSet(const UnicodeString& pattern, ParsePosition& pos,
70 const SymbolTable* symbols,
72 len(0), capacity(START_EXTRA), list(0), bmpSet(0), buffer(0),
73 bufferCapacity(0), patLen(0), pat(NULL), strings(NULL), stringSpan(NULL),
76 if(U_SUCCESS(status)){
77 list = (UChar32*) uprv_malloc(sizeof(UChar32) * capacity);
80 status = U_MEMORY_ALLOCATION_ERROR;
82 allocateStrings(status);
83 applyPattern(pattern, pos, options, symbols, status);
89 //----------------------------------------------------------------
91 //----------------------------------------------------------------
93 UnicodeSet& UnicodeSet::applyPattern(const UnicodeString& pattern,
95 const SymbolTable* symbols,
98 applyPattern(pattern, pos, options, symbols, status);
99 if (U_FAILURE(status)) return *this;
101 int32_t i = pos.getIndex();
103 if (options & USET_IGNORE_SPACE) {
104 // Skip over trailing whitespace
105 ICU_Utility::skipWhitespace(pattern, i, TRUE);
108 if (i != pattern.length()) {
109 status = U_ILLEGAL_ARGUMENT_ERROR;
114 UnicodeSet& UnicodeSet::applyPattern(const UnicodeString& pattern,
117 const SymbolTable* symbols,
118 UErrorCode& status) {
119 if (U_FAILURE(status)) {
123 status = U_NO_WRITE_PERMISSION;
126 // Need to build the pattern in a temporary string because
127 // _applyPattern calls add() etc., which set pat to empty.
128 UnicodeString rebuiltPat;
129 RuleCharacterIterator chars(pattern, symbols, pos);
130 applyPattern(chars, symbols, rebuiltPat, options, &UnicodeSet::closeOver, status);
131 if (U_FAILURE(status)) return *this;
132 if (chars.inVariable()) {
133 // syntaxError(chars, "Extra chars in variable value");
134 status = U_MALFORMED_SET;
137 setPattern(rebuiltPat);
141 // USetAdder implementation
142 // Does not use uset.h to reduce code dependencies
143 static void U_CALLCONV
144 _set_add(USet *set, UChar32 c) {
145 ((UnicodeSet *)set)->add(c);
148 static void U_CALLCONV
149 _set_addRange(USet *set, UChar32 start, UChar32 end) {
150 ((UnicodeSet *)set)->add(start, end);
153 static void U_CALLCONV
154 _set_addString(USet *set, const UChar *str, int32_t length) {
155 ((UnicodeSet *)set)->add(UnicodeString((UBool)(length<0), str, length));
158 //----------------------------------------------------------------
160 //----------------------------------------------------------------
162 // add the result of a full case mapping to the set
163 // use str as a temporary string to avoid constructing one
165 addCaseMapping(UnicodeSet &set, int32_t result, const UChar *full, UnicodeString &str) {
167 if(result > UCASE_MAX_STRING_LENGTH) {
168 // add a single-code point case mapping
171 // add a string case mapping from full with length result
172 str.setTo((UBool)FALSE, full, result);
176 // result < 0: the code point mapped to itself, no need to add it
180 UnicodeSet& UnicodeSet::closeOver(int32_t attribute) {
181 if (isFrozen() || isBogus()) {
184 if (attribute & (USET_CASE_INSENSITIVE | USET_ADD_CASE_MAPPINGS)) {
185 const UCaseProps *csp = ucase_getSingleton();
187 UnicodeSet foldSet(*this);
194 NULL, // don't need remove()
195 NULL // don't need removeRange()
198 // start with input set to guarantee inclusion
199 // USET_CASE: remove strings because the strings will actually be reduced (folded);
200 // therefore, start with no strings and add only those needed
201 if (attribute & USET_CASE_INSENSITIVE) {
202 foldSet.strings->removeAllElements();
205 int32_t n = getRangeCount();
208 int32_t locCache = 0;
210 for (int32_t i=0; i<n; ++i) {
211 UChar32 start = getRangeStart(i);
212 UChar32 end = getRangeEnd(i);
214 if (attribute & USET_CASE_INSENSITIVE) {
216 for (UChar32 cp=start; cp<=end; ++cp) {
217 ucase_addCaseClosure(csp, cp, &sa);
221 // (does not add long s for regular s, or Kelvin for k, for example)
222 for (UChar32 cp=start; cp<=end; ++cp) {
223 result = ucase_toFullLower(csp, cp, NULL, NULL, &full, "", &locCache);
224 addCaseMapping(foldSet, result, full, str);
226 result = ucase_toFullTitle(csp, cp, NULL, NULL, &full, "", &locCache);
227 addCaseMapping(foldSet, result, full, str);
229 result = ucase_toFullUpper(csp, cp, NULL, NULL, &full, "", &locCache);
230 addCaseMapping(foldSet, result, full, str);
232 result = ucase_toFullFolding(csp, cp, &full, 0);
233 addCaseMapping(foldSet, result, full, str);
237 if (strings != NULL && strings->size() > 0) {
238 if (attribute & USET_CASE_INSENSITIVE) {
239 for (int32_t j=0; j<strings->size(); ++j) {
240 str = *(const UnicodeString *) strings->elementAt(j);
242 if(!ucase_addStringCaseClosure(csp, str.getBuffer(), str.length(), &sa)) {
243 foldSet.add(str); // does not map to code points: add the folded string itself
248 #if !UCONFIG_NO_BREAK_ITERATION
249 UErrorCode status = U_ZERO_ERROR;
250 BreakIterator *bi = BreakIterator::createWordInstance(root, status);
251 if (U_SUCCESS(status)) {
253 const UnicodeString *pStr;
255 for (int32_t j=0; j<strings->size(); ++j) {
256 pStr = (const UnicodeString *) strings->elementAt(j);
257 (str = *pStr).toLower(root);
259 #if !UCONFIG_NO_BREAK_ITERATION
260 (str = *pStr).toTitle(bi, root);
263 (str = *pStr).toUpper(root);
265 (str = *pStr).foldCase();
268 #if !UCONFIG_NO_BREAK_ITERATION