1 // Copyright (C) 2016 and later: Unicode, Inc. and others.
2 // License & terms of use: http://www.unicode.org/copyright.html
4 *******************************************************************************
6 * Copyright (C) 2011, International Business Machines
7 * Corporation and others. All Rights Reserved.
9 *******************************************************************************
10 * file name: uniset_closure.cpp
12 * tab size: 8 (not used)
15 * created on: 2011may30
16 * created by: Markus W. Scherer
18 * UnicodeSet::closeOver() and related methods moved here from uniset_props.cpp
19 * to simplify dependencies.
20 * In particular, this depends on the BreakIterator, but the BreakIterator
21 * code also builds UnicodeSets from patterns and needs uniset_props.
24 #include "unicode/brkiter.h"
25 #include "unicode/locid.h"
26 #include "unicode/parsepos.h"
27 #include "unicode/uniset.h"
34 // initial storage. Must be >= 0
35 // *** same as in uniset.cpp ! ***
36 #define START_EXTRA 16
40 // TODO memory debugging provided inside uniset.cpp
41 // could be made available here but probably obsolete with use of modern
42 // memory leak checker tools
45 //----------------------------------------------------------------
47 //----------------------------------------------------------------
49 UnicodeSet::UnicodeSet(const UnicodeString& pattern,
51 const SymbolTable* symbols,
53 len(0), capacity(START_EXTRA), list(0), bmpSet(0), buffer(0),
54 bufferCapacity(0), patLen(0), pat(NULL), strings(NULL), stringSpan(NULL),
57 if(U_SUCCESS(status)){
58 list = (UChar32*) uprv_malloc(sizeof(UChar32) * capacity);
61 status = U_MEMORY_ALLOCATION_ERROR;
63 allocateStrings(status);
64 applyPattern(pattern, options, symbols, status);
70 UnicodeSet::UnicodeSet(const UnicodeString& pattern, ParsePosition& pos,
72 const SymbolTable* symbols,
74 len(0), capacity(START_EXTRA), list(0), bmpSet(0), buffer(0),
75 bufferCapacity(0), patLen(0), pat(NULL), strings(NULL), stringSpan(NULL),
78 if(U_SUCCESS(status)){
79 list = (UChar32*) uprv_malloc(sizeof(UChar32) * capacity);
82 status = U_MEMORY_ALLOCATION_ERROR;
84 allocateStrings(status);
85 applyPattern(pattern, pos, options, symbols, status);
91 //----------------------------------------------------------------
93 //----------------------------------------------------------------
95 UnicodeSet& UnicodeSet::applyPattern(const UnicodeString& pattern,
97 const SymbolTable* symbols,
100 applyPattern(pattern, pos, options, symbols, status);
101 if (U_FAILURE(status)) return *this;
103 int32_t i = pos.getIndex();
105 if (options & USET_IGNORE_SPACE) {
106 // Skip over trailing whitespace
107 ICU_Utility::skipWhitespace(pattern, i, TRUE);
110 if (i != pattern.length()) {
111 status = U_ILLEGAL_ARGUMENT_ERROR;
116 UnicodeSet& UnicodeSet::applyPattern(const UnicodeString& pattern,
119 const SymbolTable* symbols,
120 UErrorCode& status) {
121 if (U_FAILURE(status)) {
125 status = U_NO_WRITE_PERMISSION;
128 // Need to build the pattern in a temporary string because
129 // _applyPattern calls add() etc., which set pat to empty.
130 UnicodeString rebuiltPat;
131 RuleCharacterIterator chars(pattern, symbols, pos);
132 applyPattern(chars, symbols, rebuiltPat, options, &UnicodeSet::closeOver, status);
133 if (U_FAILURE(status)) return *this;
134 if (chars.inVariable()) {
135 // syntaxError(chars, "Extra chars in variable value");
136 status = U_MALFORMED_SET;
139 setPattern(rebuiltPat);
143 // USetAdder implementation
144 // Does not use uset.h to reduce code dependencies
145 static void U_CALLCONV
146 _set_add(USet *set, UChar32 c) {
147 ((UnicodeSet *)set)->add(c);
150 static void U_CALLCONV
151 _set_addRange(USet *set, UChar32 start, UChar32 end) {
152 ((UnicodeSet *)set)->add(start, end);
155 static void U_CALLCONV
156 _set_addString(USet *set, const UChar *str, int32_t length) {
157 ((UnicodeSet *)set)->add(UnicodeString((UBool)(length<0), str, length));
160 //----------------------------------------------------------------
162 //----------------------------------------------------------------
164 // add the result of a full case mapping to the set
165 // use str as a temporary string to avoid constructing one
167 addCaseMapping(UnicodeSet &set, int32_t result, const UChar *full, UnicodeString &str) {
169 if(result > UCASE_MAX_STRING_LENGTH) {
170 // add a single-code point case mapping
173 // add a string case mapping from full with length result
174 str.setTo((UBool)FALSE, full, result);
178 // result < 0: the code point mapped to itself, no need to add it
182 UnicodeSet& UnicodeSet::closeOver(int32_t attribute) {
183 if (isFrozen() || isBogus()) {
186 if (attribute & (USET_CASE_INSENSITIVE | USET_ADD_CASE_MAPPINGS)) {
187 const UCaseProps *csp = ucase_getSingleton();
189 UnicodeSet foldSet(*this);
196 NULL, // don't need remove()
197 NULL // don't need removeRange()
200 // start with input set to guarantee inclusion
201 // USET_CASE: remove strings because the strings will actually be reduced (folded);
202 // therefore, start with no strings and add only those needed
203 if (attribute & USET_CASE_INSENSITIVE) {
204 foldSet.strings->removeAllElements();
207 int32_t n = getRangeCount();
210 int32_t locCache = 0;
212 for (int32_t i=0; i<n; ++i) {
213 UChar32 start = getRangeStart(i);
214 UChar32 end = getRangeEnd(i);
216 if (attribute & USET_CASE_INSENSITIVE) {
218 for (UChar32 cp=start; cp<=end; ++cp) {
219 ucase_addCaseClosure(csp, cp, &sa);
223 // (does not add long s for regular s, or Kelvin for k, for example)
224 for (UChar32 cp=start; cp<=end; ++cp) {
225 result = ucase_toFullLower(csp, cp, NULL, NULL, &full, "", &locCache);
226 addCaseMapping(foldSet, result, full, str);
228 result = ucase_toFullTitle(csp, cp, NULL, NULL, &full, "", &locCache);
229 addCaseMapping(foldSet, result, full, str);
231 result = ucase_toFullUpper(csp, cp, NULL, NULL, &full, "", &locCache);
232 addCaseMapping(foldSet, result, full, str);
234 result = ucase_toFullFolding(csp, cp, &full, 0);
235 addCaseMapping(foldSet, result, full, str);
239 if (strings != NULL && strings->size() > 0) {
240 if (attribute & USET_CASE_INSENSITIVE) {
241 for (int32_t j=0; j<strings->size(); ++j) {
242 str = *(const UnicodeString *) strings->elementAt(j);
244 if(!ucase_addStringCaseClosure(csp, str.getBuffer(), str.length(), &sa)) {
245 foldSet.add(str); // does not map to code points: add the folded string itself
250 #if !UCONFIG_NO_BREAK_ITERATION
251 UErrorCode status = U_ZERO_ERROR;
252 BreakIterator *bi = BreakIterator::createWordInstance(root, status);
253 if (U_SUCCESS(status)) {
255 const UnicodeString *pStr;
257 for (int32_t j=0; j<strings->size(); ++j) {
258 pStr = (const UnicodeString *) strings->elementAt(j);
259 (str = *pStr).toLower(root);
261 #if !UCONFIG_NO_BREAK_ITERATION
262 (str = *pStr).toTitle(bi, root);
265 (str = *pStr).toUpper(root);
267 (str = *pStr).foldCase();
270 #if !UCONFIG_NO_BREAK_ITERATION