xref: /trunk/main/i18npool/source/collator/collator_unicode.cxx (revision 91144cd0085a7583d2099b982122deb2184ab956)
1 /**************************************************************
2  *
3  * Licensed to the Apache Software Foundation (ASF) under one
4  * or more contributor license agreements.  See the NOTICE file
5  * distributed with this work for additional information
6  * regarding copyright ownership.  The ASF licenses this file
7  * to you under the Apache License, Version 2.0 (the
8  * "License"); you may not use this file except in compliance
9  * with the License.  You may obtain a copy of the License at
10  *
11  *   http://www.apache.org/licenses/LICENSE-2.0
12  *
13  * Unless required by applicable law or agreed to in writing,
14  * software distributed under the License is distributed on an
15  * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16  * KIND, either express or implied.  See the License for the
17  * specific language governing permissions and limitations
18  * under the License.
19  *
20  *************************************************************/
21 
22 
23 
24 // MARKER(update_precomp.py): autogen include statement, do not remove
25 #include "precompiled_i18npool.hxx"
26 
27 // generated list of languages
28 #include "lrl_include.hxx"
29 
30 #include <rtl/ustrbuf.hxx>
31 #include <collator_unicode.hxx>
32 #include <localedata.hxx>
33 #include <com/sun/star/i18n/CollatorOptions.hpp>
34 
35 using namespace ::com::sun::star;
36 using namespace ::com::sun::star::lang;
37 using namespace ::com::sun::star::uno;
38 using namespace ::rtl;
39 
40 namespace com { namespace sun { namespace star { namespace i18n {
41 
Collator_Unicode()42 Collator_Unicode::Collator_Unicode()
43 {
44     implementationName = "com.sun.star.i18n.Collator_Unicode";
45     collator = NULL;
46     uca_base = NULL;
47     hModule = NULL;
48 }
49 
~Collator_Unicode()50 Collator_Unicode::~Collator_Unicode()
51 {
52     if (collator) delete collator;
53     if (uca_base) delete uca_base;
54     if (hModule) osl_unloadModule(hModule);
55 }
56 
57 sal_Int32 SAL_CALL
compareSubstring(const OUString & str1,sal_Int32 off1,sal_Int32 len1,const OUString & str2,sal_Int32 off2,sal_Int32 len2)58 Collator_Unicode::compareSubstring( const OUString& str1, sal_Int32 off1, sal_Int32 len1,
59     const OUString& str2, sal_Int32 off2, sal_Int32 len2)
60 {
61     return collator->compare(reinterpret_cast<const UChar *>(str1.getStr()) + off1, len1, reinterpret_cast<const UChar *>(str2.getStr()) + off2, len2); // UChar != sal_Unicode in MinGW
62 }
63 
64 sal_Int32 SAL_CALL
compareString(const OUString & str1,const OUString & str2)65 Collator_Unicode::compareString( const OUString& str1, const OUString& str2)
66 {
67     return collator->compare(reinterpret_cast<const UChar *>(str1.getStr()), reinterpret_cast<const UChar *>(str2.getStr()));   // UChar != sal_Unicode in MinGW
68 }
69 
thisModule()70 extern "C" { static void SAL_CALL thisModule() {} }
71 
72 sal_Int32 SAL_CALL
loadCollatorAlgorithm(const OUString & rAlgorithm,const lang::Locale & rLocale,sal_Int32 options)73 Collator_Unicode::loadCollatorAlgorithm(const OUString& rAlgorithm, const lang::Locale& rLocale, sal_Int32 options)
74 {
75     if (!collator) {
76         UErrorCode status = U_ZERO_ERROR;
77         OUString rule = LocaleData().getCollatorRuleByAlgorithm(rLocale, rAlgorithm);
78         if (rule.getLength() > 0) {
79             collator = new RuleBasedCollator(reinterpret_cast<const UChar *>(rule.getStr()), status);   // UChar != sal_Unicode in MinGW
80             if (! U_SUCCESS(status)) throw RuntimeException();
81         }
82         if (!collator && OUString::createFromAscii(LOCAL_RULE_LANGS).indexOf(rLocale.Language) >= 0) {
83             OUStringBuffer aBuf;
84 #ifdef SAL_DLLPREFIX
85             aBuf.appendAscii(SAL_DLLPREFIX);
86 #endif
87             aBuf.appendAscii( "collator_data" ).appendAscii( SAL_DLLEXTENSION );
88             hModule = osl_loadModuleRelative( &thisModule, aBuf.makeStringAndClear().pData, SAL_LOADMODULE_DEFAULT );
89             if (hModule) {
90                 const sal_uInt8* (*func)() = NULL;
91                 aBuf.appendAscii("get_").append(rLocale.Language).appendAscii("_");
92                 if (rLocale.Language.equalsAscii("zh")) {
93                     OUString func_base = aBuf.makeStringAndClear();
94                     if (OUString::createFromAscii("TW HK MO").indexOf(rLocale.Country) >= 0)
95                         func=(const sal_uInt8* (*)()) osl_getFunctionSymbol(hModule,
96                                     (func_base + OUString::createFromAscii("TW_") + rAlgorithm).pData);
97                     if (!func)
98                         func=(const sal_uInt8* (*)()) osl_getFunctionSymbol(hModule, (func_base + rAlgorithm).pData);
99                 } else {
100                     if (rLocale.Language.equalsAscii("ja")) {
101                         // replace algrithm name to implementation name.
102                         if (rAlgorithm.equalsAscii("phonetic (alphanumeric first)") )
103                             aBuf.appendAscii("phonetic_alphanumeric_first");
104                         else if (rAlgorithm.equalsAscii("phonetic (alphanumeric last)"))
105                             aBuf.appendAscii("phonetic_alphanumeric_last");
106                         else
107                             aBuf.append(rAlgorithm);
108                     } else {
109                         aBuf.append(rAlgorithm);
110                     }
111                     func=(const sal_uInt8* (*)()) osl_getFunctionSymbol(hModule, aBuf.makeStringAndClear().pData);
112                 }
113                 if (func) {
114                     const sal_uInt8* ruleImage=func();
115                     uca_base = new RuleBasedCollator(static_cast<UChar*>(NULL), status);
116                     if (! U_SUCCESS(status)) throw RuntimeException();
117                     collator = new RuleBasedCollator(reinterpret_cast<const uint8_t*>(ruleImage), -1, uca_base, status);
118                     if (! U_SUCCESS(status)) throw RuntimeException();
119                 }
120             }
121         }
122         if (!collator) {
123             /** ICU collators are loaded using a locale only.
124                 ICU uses Variant as collation algorithm name (like de__PHONEBOOK
125                 locale), note the empty territory (Country) designator in this special
126                 case here. The icu::Locale constructor changes the algorithm name to
127                 uppercase itself, so we don't have to bother with that.
128             */
129             icu::Locale icuLocale(
130                    OUStringToOString(rLocale.Language, RTL_TEXTENCODING_ASCII_US).getStr(),
131                    OUStringToOString(rLocale.Country, RTL_TEXTENCODING_ASCII_US).getStr(),
132                    OUStringToOString(rAlgorithm, RTL_TEXTENCODING_ASCII_US).getStr());
133             // load ICU collator
134             collator = (RuleBasedCollator*) icu::Collator::createInstance(icuLocale, status);
135             if (! U_SUCCESS(status)) throw RuntimeException();
136         }
137     }
138 
139     if (options & CollatorOptions::CollatorOptions_IGNORE_CASE_ACCENT)
140         collator->setStrength(Collator::PRIMARY);
141     else if (options & CollatorOptions::CollatorOptions_IGNORE_CASE)
142         collator->setStrength(Collator::SECONDARY);
143     else
144         collator->setStrength(Collator::TERTIARY);
145 
146     return(0);
147 }
148 
149 
150 OUString SAL_CALL
getImplementationName()151 Collator_Unicode::getImplementationName()
152 {
153     return OUString::createFromAscii(implementationName);
154 }
155 
156 sal_Bool SAL_CALL
supportsService(const rtl::OUString & rServiceName)157 Collator_Unicode::supportsService(const rtl::OUString& rServiceName)
158 {
159     return !rServiceName.compareToAscii(implementationName);
160 }
161 
162 Sequence< OUString > SAL_CALL
getSupportedServiceNames()163 Collator_Unicode::getSupportedServiceNames()
164 {
165     Sequence< OUString > aRet(1);
166     aRet[0] = OUString::createFromAscii(implementationName);
167     return aRet;
168 }
169 
170 } } } }
171