diff options
Diffstat (limited to 'lingucomponent/source/hyphenator')
5 files changed, 1374 insertions, 0 deletions
diff --git a/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/exports.dxp b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/exports.dxp new file mode 100644 index 000000000000..b0f85bf7bebf --- /dev/null +++ b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/exports.dxp @@ -0,0 +1,3 @@ +component_getFactory +component_getImplementationEnvironment +component_writeInfo diff --git a/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hreg.cxx b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hreg.cxx new file mode 100644 index 000000000000..448ecc529bf0 --- /dev/null +++ b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hreg.cxx @@ -0,0 +1,86 @@ +/************************************************************************* + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * Copyright 2000, 2010 Oracle and/or its affiliates. + * + * OpenOffice.org - a multi-platform office productivity suite + * + * This file is part of OpenOffice.org. + * + * OpenOffice.org is free software: you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License version 3 + * only, as published by the Free Software Foundation. + * + * OpenOffice.org is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU Lesser General Public License version 3 for more details + * (a copy is included in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU Lesser General Public License + * version 3 along with OpenOffice.org. If not, see + * <http://www.openoffice.org/license.html> + * for a copy of the LGPLv3 License. + * + ************************************************************************/ + +// MARKER(update_precomp.py): autogen include statement, do not remove +#include "precompiled_lingucomponent.hxx" + + + +#include <cppuhelper/factory.hxx> // helper for factories +#include <rtl/string.hxx> + +#include <com/sun/star/registry/XRegistryKey.hpp> + +using namespace com::sun::star::lang; +using namespace com::sun::star::registry; + +//////////////////////////////////////// +// declaration of external RegEntry-functions defined by the service objects +// + +extern sal_Bool SAL_CALL Hyphenator_writeInfo( + void * /*pServiceManager*/, XRegistryKey * pRegistryKey ); + +extern void * SAL_CALL Hyphenator_getFactory( + const sal_Char * pImplName, + XMultiServiceFactory * pServiceManager, + void * /*pRegistryKey*/ ); + +//////////////////////////////////////// +// definition of the two functions that are used to provide the services +// + +extern "C" +{ + +sal_Bool SAL_CALL component_writeInfo( + void * pServiceManager, XRegistryKey * pRegistryKey ) +{ + return Hyphenator_writeInfo( pServiceManager, pRegistryKey ); +} + +void SAL_CALL component_getImplementationEnvironment( + const sal_Char ** ppEnvTypeName, uno_Environment ** /*ppEnv*/ ) +{ + *ppEnvTypeName = CPPU_CURRENT_LANGUAGE_BINDING_NAME; +} + +void * SAL_CALL component_getFactory( + const sal_Char * pImplName, void * pServiceManager, void * pRegistryKey ) +{ + void * pRet = Hyphenator_getFactory( + pImplName, + reinterpret_cast< XMultiServiceFactory * >( pServiceManager ), + pRegistryKey ); + + return pRet; +} + +} + +/////////////////////////////////////////////////////////////////////////// + diff --git a/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hyphenimp.cxx b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hyphenimp.cxx new file mode 100755 index 000000000000..3b0871cc36b6 --- /dev/null +++ b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hyphenimp.cxx @@ -0,0 +1,982 @@ +/************************************************************************* + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * Copyright 2000, 2010 Oracle and/or its affiliates. + * + * OpenOffice.org - a multi-platform office productivity suite + * + * This file is part of OpenOffice.org. + * + * OpenOffice.org is free software: you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License version 3 + * only, as published by the Free Software Foundation. + * + * OpenOffice.org is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU Lesser General Public License version 3 for more details + * (a copy is included in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU Lesser General Public License + * version 3 along with OpenOffice.org. If not, see + * <http://www.openoffice.org/license.html> + * for a copy of the LGPLv3 License. + * + ************************************************************************/ + +// MARKER(update_precomp.py): autogen include statement, do not remove +#include "precompiled_lingucomponent.hxx" + + +#include <com/sun/star/uno/Reference.h> +#include <com/sun/star/linguistic2/XSearchableDictionaryList.hpp> + +//#include <com/sun/star/linguistic2/SpellFailure.hpp> +#include <cppuhelper/factory.hxx> // helper for factories +#include <com/sun/star/registry/XRegistryKey.hpp> +#include <i18npool/mslangid.hxx> +#include <unotools/pathoptions.hxx> +#include <unotools/useroptions.hxx> +#include <tools/debug.hxx> +#include <unotools/processfactory.hxx> +#include <osl/mutex.hxx> + +#include <hyphen.h> + +#ifndef _HYPHENIMP_HXX +#include <hyphenimp.hxx> +#endif + +#include <linguistic/hyphdta.hxx> +#include <rtl/ustring.hxx> + +#include <rtl/ustrbuf.hxx> + +#include <linguistic/lngprops.hxx> +#include <unotools/pathoptions.hxx> +#include <unotools/useroptions.hxx> +#include <unotools/lingucfg.hxx> +#include <osl/file.hxx> + +#include "dictmgr.hxx" + +#include <stdio.h> +#include <string.h> + +#include <list> +#include <set> + +using namespace utl; +using namespace osl; +using namespace rtl; +using namespace com::sun::star; +using namespace com::sun::star::beans; +using namespace com::sun::star::lang; +using namespace com::sun::star::uno; +using namespace com::sun::star::linguistic2; +using namespace linguistic; + +// values asigned to capitalization types +#define CAPTYPE_UNKNOWN 0 +#define CAPTYPE_NOCAP 1 +#define CAPTYPE_INITCAP 2 +#define CAPTYPE_ALLCAP 3 +#define CAPTYPE_MIXED 4 + +// min, max + +//#define Min(a,b) (a < b ? a : b) +#define Max(a,b) (a > b ? a : b) + +/////////////////////////////////////////////////////////////////////////// + + +Hyphenator::Hyphenator() : + aEvtListeners ( GetLinguMutex() ) +{ + bDisposing = FALSE; + pPropHelper = NULL; + aDicts = NULL; + numdict = 0; + +} + + +Hyphenator::~Hyphenator() +{ + if (pPropHelper) + pPropHelper->RemoveAsPropListener(); + + if ((numdict) && (aDicts)) { + for (int i=0; i < numdict; i++) { + if (aDicts[i].apCC) delete aDicts[i].apCC; + aDicts[i].apCC = NULL; + } + } + if (aDicts) delete[] aDicts; + aDicts = NULL; + numdict = 0; +} + + +PropertyHelper_Hyphen & Hyphenator::GetPropHelper_Impl() +{ + + if (!pPropHelper) + { + Reference< XPropertySet > xPropSet( GetLinguProperties(), UNO_QUERY ); + + pPropHelper = new PropertyHelper_Hyphen ((XHyphenator *) this, xPropSet ); + xPropHelper = pPropHelper; + pPropHelper->AddAsPropListener(); //! after a reference is established + } + return *pPropHelper; + +} + + +Sequence< Locale > SAL_CALL Hyphenator::getLocales() + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + // this routine should return the locales supported by the installed + // dictionaries. + + if (!numdict) + { + SvtLinguConfig aLinguCfg; + + // get list of dictionaries-to-use + // (or better speaking: the list of dictionaries using the + // new configuration entries). + std::list< SvtLinguConfigDictionaryEntry > aDics; + uno::Sequence< rtl::OUString > aFormatList; + aLinguCfg.GetSupportedDictionaryFormatsFor( A2OU("Hyphenators"), + A2OU("org.openoffice.lingu.LibHnjHyphenator"), aFormatList ); + sal_Int32 nLen = aFormatList.getLength(); + for (sal_Int32 i = 0; i < nLen; ++i) + { + std::vector< SvtLinguConfigDictionaryEntry > aTmpDic( + aLinguCfg.GetActiveDictionariesByFormat( aFormatList[i] ) ); + aDics.insert( aDics.end(), aTmpDic.begin(), aTmpDic.end() ); + } + + //!! for compatibility with old dictionaries (the ones not using extensions + //!! or new configuration entries, but still using the dictionary.lst file) + //!! Get the list of old style spell checking dictionaries to use... + std::vector< SvtLinguConfigDictionaryEntry > aOldStyleDics( + GetOldStyleDics( "HYPH" ) ); + + // to prefer dictionaries with configuration entries we will only + // use those old style dictionaries that add a language that + // is not yet supported by the list od new style dictionaries + MergeNewStyleDicsAndOldStyleDics( aDics, aOldStyleDics ); + + numdict = aDics.size(); + if (numdict) + { + // get supported locales from the dictionaries-to-use... + sal_Int32 k = 0; + std::set< rtl::OUString, lt_rtl_OUString > aLocaleNamesSet; + std::list< SvtLinguConfigDictionaryEntry >::const_iterator aDictIt; + for (aDictIt = aDics.begin(); aDictIt != aDics.end(); ++aDictIt) + { + uno::Sequence< rtl::OUString > aLocaleNames( aDictIt->aLocaleNames ); + sal_Int32 nLen2 = aLocaleNames.getLength(); + for (k = 0; k < nLen2; ++k) + { + aLocaleNamesSet.insert( aLocaleNames[k] ); + } + } + // ... and add them to the resulting sequence + aSuppLocales.realloc( aLocaleNamesSet.size() ); + std::set< rtl::OUString, lt_rtl_OUString >::const_iterator aItB; + k = 0; + for (aItB = aLocaleNamesSet.begin(); aItB != aLocaleNamesSet.end(); ++aItB) + { + Locale aTmp( MsLangId::convertLanguageToLocale( + MsLangId::convertIsoStringToLanguage( *aItB ))); + aSuppLocales[k++] = aTmp; + } + + //! For each dictionary and each locale we need a seperate entry. + //! If this results in more than one dictionary per locale than (for now) + //! it is undefined which dictionary gets used. + //! In the future the implementation should support using several dictionaries + //! for one locale. + numdict = 0; + for (aDictIt = aDics.begin(); aDictIt != aDics.end(); ++aDictIt) + numdict = numdict + aDictIt->aLocaleNames.getLength(); + + // add dictionary information + aDicts = new HDInfo[numdict]; +/* + aTEncs = new rtl_TextEncoding [numdict]; + aTLocs = new Locale [numdict]; + aTNames = new OUString [numdict]; + aCharSetInfo = new CharClass* [numdict]; +*/ + k = 0; + for (aDictIt = aDics.begin(); aDictIt != aDics.end(); ++aDictIt) + { + if (aDictIt->aLocaleNames.getLength() > 0 && + aDictIt->aLocations.getLength() > 0) + { + uno::Sequence< rtl::OUString > aLocaleNames( aDictIt->aLocaleNames ); + sal_Int32 nLocales = aLocaleNames.getLength(); + + // currently only one language per dictionary is supported in the actual implementation... + // Thus here we work-around this by adding the same dictionary several times. + // Once for each of it's supported locales. + for (sal_Int32 i = 0; i < nLocales; ++i) + { + aDicts[k].aPtr = NULL; + aDicts[k].aEnc = 0; + aDicts[k].aLoc = MsLangId::convertLanguageToLocale( + MsLangId::convertIsoStringToLanguage( aDictIt->aLocaleNames[i] )); + aDicts[k].apCC = new CharClass( aDicts[k].aLoc ); + // also both files have to be in the same directory and the + // file names must only differ in the extension (.aff/.dic). + // Thus we use the first location only and strip the extension part. + rtl::OUString aLocation = aDictIt->aLocations[0]; + sal_Int32 nPos = aLocation.lastIndexOf( '.' ); + aLocation = aLocation.copy( 0, nPos ); + aDicts[k].aName = aLocation; + + ++k; + } + } + } + DBG_ASSERT( k == numdict, "index mismatch?" ); + } + else + { + /* no dictionary found so register no dictionaries */ + numdict = 0; + aDicts = NULL; + aSuppLocales.realloc(0); + } + } + + return aSuppLocales; +} + + + +sal_Bool SAL_CALL Hyphenator::hasLocale(const Locale& rLocale) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + BOOL bRes = FALSE; + if (!aSuppLocales.getLength()) + getLocales(); + + const Locale *pLocale = aSuppLocales.getConstArray(); + INT32 nLen = aSuppLocales.getLength(); + for (INT32 i = 0; i < nLen; ++i) + { + if (rLocale == pLocale[i]) + { + bRes = TRUE; + break; + } + } + return bRes; +} + + +Reference< XHyphenatedWord > SAL_CALL +Hyphenator::hyphenate( const ::rtl::OUString& aWord, + const ::com::sun::star::lang::Locale& aLocale, + sal_Int16 nMaxLeading, + const ::com::sun::star::beans::PropertyValues& aProperties ) + throw (com::sun::star::uno::RuntimeException, + com::sun::star::lang::IllegalArgumentException) +{ + int nHyphenationPos = -1; + int nHyphenationPosAlt = -1; + int nHyphenationPosAltHyph = -1; + int wordlen; + char *hyphens; + char *lcword; + int k = 0; + + PropertyHelper_Hyphen & rHelper = GetPropHelper(); + rHelper.SetTmpPropVals(aProperties); + sal_Int16 minTrail = rHelper.GetMinTrailing(); + sal_Int16 minLead = rHelper.GetMinLeading(); + sal_Int16 minLen = rHelper.GetMinWordLength(); + + HyphenDict *dict = NULL; + rtl_TextEncoding aEnc = 0; + CharClass * pCC = NULL; + + Reference< XHyphenatedWord > xRes; + + k = -1; + for (int j = 0; j < numdict; j++) + if (aLocale == aDicts[j].aLoc) k = j; + + + // if we have a hyphenation dictionary matching this locale + if (k != -1) { + + // if this dictinary has not been loaded yet do that + if (!aDicts[k].aPtr) { + + OUString DictFN = aDicts[k].aName + A2OU(".dic"); + OUString dictpath; + + osl::FileBase::getSystemPathFromFileURL( DictFN, dictpath ); + OString sTmp( OU2ENC( dictpath, osl_getThreadTextEncoding() ) ); + +#if defined(WNT) + // workaround for Windows specifc problem that the + // path length in calls to 'fopen' is limted to somewhat + // about 120+ characters which will usually be exceed when + // using dictionaries as extensions. + sTmp = Win_GetShortPathName( dictpath ); +#endif + + if ( ( dict = hnj_hyphen_load ( sTmp.getStr()) ) == NULL ) + { + fprintf(stderr, "Couldn't find file %s\n", OU2ENC(dictpath, osl_getThreadTextEncoding()) ); + return NULL; + } + aDicts[k].aPtr = dict; + aDicts[k].aEnc = rtl_getTextEncodingFromUnixCharset(dict->cset); + if (aDicts[k].aEnc == RTL_TEXTENCODING_DONTKNOW) { + if (strcmp("ISCII-DEVANAGARI", dict->cset) == 0) { + aDicts[k].aEnc = RTL_TEXTENCODING_ISCII_DEVANAGARI; + } else if (strcmp("UTF-8", dict->cset) == 0) { + aDicts[k].aEnc = RTL_TEXTENCODING_UTF8; + } + } + } + + // other wise hyphenate the word with that dictionary + dict = aDicts[k].aPtr; + aEnc = aDicts[k].aEnc; + pCC = aDicts[k].apCC; + + sal_uInt16 ct = CAPTYPE_UNKNOWN; + ct = capitalType(aWord, pCC); + + // first convert any smart quotes or apostrophes to normal ones + OUStringBuffer rBuf(aWord); + sal_Int32 nc = rBuf.getLength(); + sal_Unicode ch; + for (sal_Int32 ix=0; ix < nc; ix++) { + ch = rBuf.charAt(ix); + if ((ch == 0x201C) || (ch == 0x201D)) rBuf.setCharAt(ix,(sal_Unicode)0x0022); + if ((ch == 0x2018) || (ch == 0x2019)) rBuf.setCharAt(ix,(sal_Unicode)0x0027); + } + OUString nWord(rBuf.makeStringAndClear()); + + // now convert word to all lowercase for pattern recognition + OUString nTerm(makeLowerCase(nWord, pCC)); + + // now convert word to needed encoding + OString encWord(OU2ENC(nTerm,aEnc)); + + wordlen = encWord.getLength(); + lcword = new char[wordlen + 1]; + hyphens = new char[wordlen + 5]; + + char ** rep = NULL; // replacements of discretionary hyphenation + int * pos = NULL; // array of [hyphenation point] minus [deletion position] + int * cut = NULL; // length of deletions in original word + + // copy converted word into simple char buffer + strcpy(lcword,encWord.getStr()); + + // now strip off any ending periods + int n = wordlen-1; + while((n >=0) && (lcword[n] == '.')) n--; + n++; + if (n > 0) { + if (hnj_hyphen_hyphenate3(dict, lcword, n, hyphens, NULL, &rep, &pos, &cut, + minLead, minTrail, Max(dict->clhmin, Max(dict->clhmin, 2) + Max(0, minLead - Max(dict->lhmin, 2))), + Max(dict->crhmin, Max(dict->crhmin, 2) + Max(0, minTrail - Max(dict->rhmin, 2))))) + { + //whoops something did not work + delete[] hyphens; + delete[] lcword; + if (rep) { + for(int j = 0; j < n; j++) { + if (rep[j]) free(rep[j]); + } + free(rep); + } + if (pos) free(pos); + if (cut) free(cut); + return NULL; + } + } + + // now backfill hyphens[] for any removed trailing periods + for (int c = n; c < wordlen; c++) hyphens[c] = '0'; + hyphens[wordlen] = '\0'; + + INT32 Leading = GetPosInWordToCheck( aWord, nMaxLeading ); + + for (INT32 i = 0; i < n; i++) + { + int leftrep = 0; + BOOL hit = (n >= minLen); + if (!rep || !rep[i] || (i >= n)) { + hit = hit && (hyphens[i]&1) && (i < Leading); + hit = hit && (i >= (minLead-1) ); + hit = hit && ((n - i - 1) >= minTrail); + } else { + // calculate change character length before hyphenation point signed with '=' + for (char * c = rep[i]; *c && (*c != '='); c++) { + if (aEnc == RTL_TEXTENCODING_UTF8) { + if (((unsigned char) *c) >> 6 != 2) leftrep++; + } else leftrep++; + } + hit = hit && (hyphens[i]&1) && ((i + leftrep - pos[i]) < Leading); + hit = hit && ((i + leftrep - pos[i]) >= (minLead-1) ); + hit = hit && ((n - i - 1 + sal::static_int_cast< sal_sSize >(strlen(rep[i])) - leftrep - 1) >= minTrail); + } + if (hit) { + nHyphenationPos = i; + if (rep && (i < n) && rep[i]) { + nHyphenationPosAlt = i - pos[i]; + nHyphenationPosAltHyph = i + leftrep - pos[i]; + } + } + } + + if (nHyphenationPos == -1) { + xRes = NULL; + } else { + if (rep && rep[nHyphenationPos]) { + // remove equal sign + char * s = rep[nHyphenationPos]; + int eq = 0; + for (; *s; s++) { + if (*s == '=') eq = 1; + if (eq) *s = *(s + 1); + } + OUString repHyphlow(rep[nHyphenationPos], strlen(rep[nHyphenationPos]), aEnc); + OUString repHyph; + switch (ct) { + case CAPTYPE_ALLCAP: + { + repHyph = makeUpperCase(repHyphlow, pCC); + break; + } + case CAPTYPE_INITCAP: + { + if (nHyphenationPosAlt == 0) { + repHyph = makeInitCap(repHyphlow, pCC); + } else { + repHyph = repHyphlow; + } + break; + } + default: + { + repHyph = repHyphlow; + break; + } + } + + // handle shortening + INT16 nPos = (INT16) ((nHyphenationPosAltHyph < nHyphenationPos) ? + nHyphenationPosAltHyph : nHyphenationPos); + // dicretionary hyphenation + xRes = new HyphenatedWord( aWord, LocaleToLanguage( aLocale ), nPos, + aWord.replaceAt(nHyphenationPosAlt + 1, cut[nHyphenationPos], repHyph), + (INT16) nHyphenationPosAltHyph); + } else { + xRes = new HyphenatedWord( aWord, LocaleToLanguage( aLocale ), + (INT16)nHyphenationPos, aWord, (INT16) nHyphenationPos); + } + } + + delete[] lcword; + delete[] hyphens; + if (rep) { + for(int j = 0; j < n; j++) { + if (rep[j]) free(rep[j]); + } + free(rep); + } + if (pos) free(pos); + if (cut) free(cut); + return xRes; + } + return NULL; +} + + +Reference < XHyphenatedWord > SAL_CALL + Hyphenator::queryAlternativeSpelling( const ::rtl::OUString& /*aWord*/, + const ::com::sun::star::lang::Locale& /*aLocale*/, + sal_Int16 /*nIndex*/, + const ::com::sun::star::beans::PropertyValues& /*aProperties*/ ) + throw(::com::sun::star::lang::IllegalArgumentException, + ::com::sun::star::uno::RuntimeException) +{ + /* alternative spelling isn't supported by tex dictionaries */ + /* XXX: OOo's extended libhjn algorithm can support alternative spellings with extended TeX dic. */ + /* TASK: implement queryAlternativeSpelling() */ + return NULL; +} + +Reference< XPossibleHyphens > SAL_CALL + Hyphenator::createPossibleHyphens( const ::rtl::OUString& aWord, + const ::com::sun::star::lang::Locale& aLocale, + const ::com::sun::star::beans::PropertyValues& aProperties ) + throw(::com::sun::star::lang::IllegalArgumentException, + ::com::sun::star::uno::RuntimeException) + +{ + int wordlen; + char *hyphens; + char *lcword; + int k; + + PropertyHelper_Hyphen & rHelper = GetPropHelper(); + rHelper.SetTmpPropVals(aProperties); + sal_Int16 minTrail = rHelper.GetMinTrailing(); + sal_Int16 minLead = rHelper.GetMinLeading(); + + HyphenDict *dict = NULL; + rtl_TextEncoding aEnc = 0; + CharClass* pCC = NULL; + + Reference< XPossibleHyphens > xRes; + + k = -1; + for (int j = 0; j < numdict; j++) + if (aLocale == aDicts[j].aLoc) k = j; + + + // if we have a hyphenation dictionary matching this locale + if (k != -1) { + + // if this dictioanry has not been loaded yet do that + if (!aDicts[k].aPtr) { + + OUString DictFN = aDicts[k].aName + A2OU(".dic"); + OUString dictpath; + + osl::FileBase::getSystemPathFromFileURL( DictFN, dictpath ); + OString sTmp( OU2ENC( dictpath, osl_getThreadTextEncoding() ) ); + +#if defined(WNT) + // workaround for Windows specifc problem that the + // path length in calls to 'fopen' is limted to somewhat + // about 120+ characters which will usually be exceed when + // using dictionaries as extensions. + sTmp = Win_GetShortPathName( dictpath ); +#endif + + if ( ( dict = hnj_hyphen_load ( sTmp.getStr()) ) == NULL ) + { + fprintf(stderr, "Couldn't find file %s and %s\n", sTmp.getStr(), OU2ENC(dictpath, osl_getThreadTextEncoding()) ); + return NULL; + } + aDicts[k].aPtr = dict; + aDicts[k].aEnc = rtl_getTextEncodingFromUnixCharset(dict->cset); + if (aDicts[k].aEnc == RTL_TEXTENCODING_DONTKNOW) { + if (strcmp("ISCII-DEVANAGARI", dict->cset) == 0) { + aDicts[k].aEnc = RTL_TEXTENCODING_ISCII_DEVANAGARI; + } else if (strcmp("UTF-8", dict->cset) == 0) { + aDicts[k].aEnc = RTL_TEXTENCODING_UTF8; + } + } + } + + // other wise hyphenate the word with that dictionary + dict = aDicts[k].aPtr; + aEnc = aDicts[k].aEnc; + pCC = aDicts[k].apCC; + + // first handle smart quotes both single and double + OUStringBuffer rBuf(aWord); + sal_Int32 nc = rBuf.getLength(); + sal_Unicode ch; + for (sal_Int32 ix=0; ix < nc; ix++) { + ch = rBuf.charAt(ix); + if ((ch == 0x201C) || (ch == 0x201D)) rBuf.setCharAt(ix,(sal_Unicode)0x0022); + if ((ch == 0x2018) || (ch == 0x2019)) rBuf.setCharAt(ix,(sal_Unicode)0x0027); + } + OUString nWord(rBuf.makeStringAndClear()); + + // now convert word to all lowercase for pattern recognition + OUString nTerm(makeLowerCase(nWord, pCC)); + + // now convert word to needed encoding + OString encWord(OU2ENC(nTerm,aEnc)); + + wordlen = encWord.getLength(); + lcword = new char[wordlen+1]; + hyphens = new char[wordlen+5]; + char ** rep = NULL; // replacements of discretionary hyphenation + int * pos = NULL; // array of [hyphenation point] minus [deletion position] + int * cut = NULL; // length of deletions in original word + + // copy converted word into simple char buffer + strcpy(lcword,encWord.getStr()); + + // first remove any trailing periods + int n = wordlen-1; + while((n >=0) && (lcword[n] == '.')) n--; + n++; + // fprintf(stderr,"hyphenate... %s\n",lcword); fflush(stderr); + if (n > 0) { + if (hnj_hyphen_hyphenate3(dict, lcword, n, hyphens, NULL, &rep, &pos, &cut, + minLead, minTrail, Max(dict->clhmin, Max(dict->clhmin, 2) + Max(0, minLead - Max(dict->lhmin, 2))), + Max(dict->crhmin, Max(dict->crhmin, 2) + Max(0, minTrail - Max(dict->rhmin, 2))))) + { + delete[] hyphens; + delete[] lcword; + + if (rep) { + for(int j = 0; j < n; j++) { + if (rep[j]) free(rep[j]); + } + free(rep); + } + if (pos) free(pos); + if (cut) free(cut); + + return NULL; + } + } + // now backfill hyphens[] for any removed periods + for (int c = n; c < wordlen; c++) hyphens[c] = '0'; + hyphens[wordlen] = '\0'; + // fprintf(stderr,"... %s\n",hyphens); fflush(stderr); + + INT16 nHyphCount = 0; + INT16 i; + + for ( i = 0; i < encWord.getLength(); i++) + if (hyphens[i]&1 && (!rep || !rep[i])) + nHyphCount++; + + Sequence< INT16 > aHyphPos(nHyphCount); + INT16 *pPos = aHyphPos.getArray(); + OUStringBuffer hyphenatedWordBuffer; + OUString hyphenatedWord; + nHyphCount = 0; + + for (i = 0; i < nWord.getLength(); i++) { + hyphenatedWordBuffer.append(aWord[i]); + // hyphenation position (not alternative) + if (hyphens[i]&1 && (!rep || !rep[i])) { + pPos[nHyphCount] = i; + hyphenatedWordBuffer.append(sal_Unicode('=')); + nHyphCount++; + } + } + + hyphenatedWord = hyphenatedWordBuffer.makeStringAndClear(); + //fprintf(stderr,"result is %s\n",OU2A(hyphenatedWord)); + //fflush(stderr); + + xRes = new PossibleHyphens( aWord, LocaleToLanguage( aLocale ), + hyphenatedWord, aHyphPos ); + + delete[] hyphens; + delete[] lcword; + + if (rep) { + for(int j = 0; j < n; j++) { + if (rep[j]) free(rep[j]); + } + free(rep); + } + if (pos) free(pos); + if (cut) free(cut); + + return xRes; + } + + return NULL; + +} + +sal_uInt16 SAL_CALL Hyphenator::capitalType(const OUString& aTerm, CharClass * pCC) +{ + sal_Int32 tlen = aTerm.getLength(); + if ((pCC) && (tlen)) { + String aStr(aTerm); + sal_Int32 nc = 0; + for (xub_StrLen tindex = 0; tindex < tlen; tindex++) { + if (pCC->getCharacterType(aStr,tindex) & + ::com::sun::star::i18n::KCharacterType::UPPER) nc++; + } + + if (nc == 0) return (sal_uInt16) CAPTYPE_NOCAP; + + if (nc == tlen) return (sal_uInt16) CAPTYPE_ALLCAP; + + if ((nc == 1) && (pCC->getCharacterType(aStr,0) & + ::com::sun::star::i18n::KCharacterType::UPPER)) + return (sal_uInt16) CAPTYPE_INITCAP; + + return (sal_uInt16) CAPTYPE_MIXED; + } + return (sal_uInt16) CAPTYPE_UNKNOWN; +} + +OUString SAL_CALL Hyphenator::makeLowerCase(const OUString& aTerm, CharClass * pCC) +{ + if (pCC) + return pCC->toLower_rtl(aTerm, 0, aTerm.getLength()); + return aTerm; +} + +OUString SAL_CALL Hyphenator::makeUpperCase(const OUString& aTerm, CharClass * pCC) +{ + if (pCC) + return pCC->toUpper_rtl(aTerm, 0, aTerm.getLength()); + return aTerm; +} + + +OUString SAL_CALL Hyphenator::makeInitCap(const OUString& aTerm, CharClass * pCC) +{ + sal_Int32 tlen = aTerm.getLength(); + if ((pCC) && (tlen)) { + OUString bTemp = aTerm.copy(0,1); + if (tlen > 1) + return ( pCC->toUpper_rtl(bTemp, 0, 1) + + pCC->toLower_rtl(aTerm,1,(tlen-1)) ); + + return pCC->toUpper_rtl(bTemp, 0, 1); + } + return aTerm; +} + + + + +Reference< XInterface > SAL_CALL Hyphenator_CreateInstance( + const Reference< XMultiServiceFactory > & /*rSMgr*/ ) + throw(Exception) +{ + + Reference< XInterface > xService = (cppu::OWeakObject*) new Hyphenator; + return xService; +} + + +sal_Bool SAL_CALL + Hyphenator::addLinguServiceEventListener( + const Reference< XLinguServiceEventListener >& rxLstnr ) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + BOOL bRes = FALSE; + if (!bDisposing && rxLstnr.is()) + { + bRes = GetPropHelper().addLinguServiceEventListener( rxLstnr ); + } + return bRes; +} + + +sal_Bool SAL_CALL + Hyphenator::removeLinguServiceEventListener( + const Reference< XLinguServiceEventListener >& rxLstnr ) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + BOOL bRes = FALSE; + if (!bDisposing && rxLstnr.is()) + { + DBG_ASSERT( xPropHelper.is(), "xPropHelper non existent" ); + bRes = GetPropHelper().removeLinguServiceEventListener( rxLstnr ); + } + return bRes; +} + + +OUString SAL_CALL + Hyphenator::getServiceDisplayName( const Locale& /*rLocale*/ ) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + return A2OU( "Libhyphen Hyphenator" ); +} + + +void SAL_CALL + Hyphenator::initialize( const Sequence< Any >& rArguments ) + throw(Exception, RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + if (!pPropHelper) + { + INT32 nLen = rArguments.getLength(); + if (2 == nLen) + { + Reference< XPropertySet > xPropSet; + rArguments.getConstArray()[0] >>= xPropSet; + //rArguments.getConstArray()[1] >>= xDicList; + + //! Pointer allows for access of the non-UNO functions. + //! And the reference to the UNO-functions while increasing + //! the ref-count and will implicitly free the memory + //! when the object is not longer used. + pPropHelper = new PropertyHelper_Hyphen( (XHyphenator *) this, xPropSet ); + xPropHelper = pPropHelper; + pPropHelper->AddAsPropListener(); //! after a reference is established + } + else { + DBG_ERROR( "wrong number of arguments in sequence" ); + } + + } +} + + +void SAL_CALL + Hyphenator::dispose() + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + if (!bDisposing) + { + bDisposing = TRUE; + EventObject aEvtObj( (XHyphenator *) this ); + aEvtListeners.disposeAndClear( aEvtObj ); + } +} + + +void SAL_CALL + Hyphenator::addEventListener( const Reference< XEventListener >& rxListener ) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + if (!bDisposing && rxListener.is()) + aEvtListeners.addInterface( rxListener ); +} + + +void SAL_CALL + Hyphenator::removeEventListener( const Reference< XEventListener >& rxListener ) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + if (!bDisposing && rxListener.is()) + aEvtListeners.removeInterface( rxListener ); +} + + +/////////////////////////////////////////////////////////////////////////// +// Service specific part +// + +OUString SAL_CALL Hyphenator::getImplementationName() + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + return getImplementationName_Static(); +} + + +sal_Bool SAL_CALL Hyphenator::supportsService( const OUString& ServiceName ) + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + Sequence< OUString > aSNL = getSupportedServiceNames(); + const OUString * pArray = aSNL.getConstArray(); + for( INT32 i = 0; i < aSNL.getLength(); i++ ) + if( pArray[i] == ServiceName ) + return TRUE; + return FALSE; +} + + +Sequence< OUString > SAL_CALL Hyphenator::getSupportedServiceNames() + throw(RuntimeException) +{ + MutexGuard aGuard( GetLinguMutex() ); + + return getSupportedServiceNames_Static(); +} + + +Sequence< OUString > Hyphenator::getSupportedServiceNames_Static() + throw() +{ + MutexGuard aGuard( GetLinguMutex() ); + + Sequence< OUString > aSNS( 1 ); // auch mehr als 1 Service moeglich + aSNS.getArray()[0] = A2OU( SN_HYPHENATOR ); + return aSNS; +} + + +sal_Bool SAL_CALL Hyphenator_writeInfo( + void * /*pServiceManager*/, registry::XRegistryKey * pRegistryKey ) +{ + + try + { + String aImpl( '/' ); + aImpl += Hyphenator::getImplementationName_Static().getStr(); + aImpl.AppendAscii( "/UNO/SERVICES" ); + Reference< registry::XRegistryKey > xNewKey = + pRegistryKey->createKey( aImpl ); + Sequence< OUString > aServices = + Hyphenator::getSupportedServiceNames_Static(); + for( INT32 i = 0; i < aServices.getLength(); i++ ) + xNewKey->createKey( aServices.getConstArray()[i] ); + + return sal_True; + } + catch(Exception &) + { + return sal_False; + } +} + + +void * SAL_CALL Hyphenator_getFactory( const sal_Char * pImplName, + XMultiServiceFactory * pServiceManager, void * ) +{ + void * pRet = 0; + if ( !Hyphenator::getImplementationName_Static().compareToAscii( pImplName ) ) + { + Reference< XSingleServiceFactory > xFactory = + cppu::createOneInstanceFactory( + pServiceManager, + Hyphenator::getImplementationName_Static(), + Hyphenator_CreateInstance, + Hyphenator::getSupportedServiceNames_Static()); + // acquire, because we return an interface pointer instead of a reference + xFactory->acquire(); + pRet = xFactory.get(); + } + return pRet; +} + + +/////////////////////////////////////////////////////////////////////////// + +#undef CAPTYPE_UNKNOWN +#undef CAPTYPE_NOCAP +#undef CAPTYPE_INITCAP +#undef CAPTYPE_ALLCAP +#undef CAPTYPE_MIXED diff --git a/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hyphenimp.hxx b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hyphenimp.hxx new file mode 100644 index 000000000000..031217914754 --- /dev/null +++ b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/hyphenimp.hxx @@ -0,0 +1,210 @@ +/************************************************************************* + * + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * Copyright 2000, 2010 Oracle and/or its affiliates. + * + * OpenOffice.org - a multi-platform office productivity suite + * + * This file is part of OpenOffice.org. + * + * OpenOffice.org is free software: you can redistribute it and/or modify + * it under the terms of the GNU Lesser General Public License version 3 + * only, as published by the Free Software Foundation. + * + * OpenOffice.org is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU Lesser General Public License version 3 for more details + * (a copy is included in the LICENSE file that accompanied this code). + * + * You should have received a copy of the GNU Lesser General Public License + * version 3 along with OpenOffice.org. If not, see + * <http://www.openoffice.org/license.html> + * for a copy of the LGPLv3 License. + * + ************************************************************************/ + + +#ifndef _LINGU2_HYPHENIMP_HXX_ +#define _LINGU2_HYPHENIMP_HXX_ + +#include <uno/lbnames.h> // CPPU_CURRENT_LANGUAGE_BINDING_NAME macro, which specify the environment type +#include <cppuhelper/implbase1.hxx> // helper for implementations +#include <cppuhelper/implbase6.hxx> // helper for implementations +#include <com/sun/star/lang/XComponent.hpp> +#include <com/sun/star/lang/XInitialization.hpp> +#include <com/sun/star/lang/XServiceDisplayName.hpp> +#include <com/sun/star/beans/XPropertySet.hpp> +#include <com/sun/star/beans/PropertyValues.hpp> +#include <com/sun/star/lang/XServiceInfo.hpp> +#include <com/sun/star/linguistic2/XHyphenator.hpp> +#include <com/sun/star/linguistic2/XSearchableDictionaryList.hpp> +#include <com/sun/star/linguistic2/XLinguServiceEventBroadcaster.hpp> +#include <tools/table.hxx> + +#include <unotools/charclass.hxx> + +#include <linguistic/misc.hxx> +#include <linguistic/lngprophelp.hxx> + +#include <lingutil.hxx> +#include <stdio.h> + +using namespace ::rtl; +using namespace ::com::sun::star::uno; +using namespace ::com::sun::star::beans; +using namespace ::com::sun::star::lang; +using namespace ::com::sun::star::linguistic2; + + +/////////////////////////////////////////////////////////////////////////// + + +struct HDInfo { + HyphenDict * aPtr; + OUString aName; + Locale aLoc; + rtl_TextEncoding aEnc; + CharClass * apCC; +}; + + + +class Hyphenator : + public cppu::WeakImplHelper6 + < + XHyphenator, + XLinguServiceEventBroadcaster, + XInitialization, + XComponent, + XServiceInfo, + XServiceDisplayName + > +{ + Sequence< Locale > aSuppLocales; + HDInfo * aDicts; + sal_Int32 numdict; + + ::cppu::OInterfaceContainerHelper aEvtListeners; + Reference< XPropertyChangeListener > xPropHelper; + Reference< XMultiServiceFactory > rSMgr; + linguistic::PropertyHelper_Hyphen * pPropHelper; + BOOL bDisposing; + + // disallow copy-constructor and assignment-operator for now + Hyphenator(const Hyphenator &); + Hyphenator & operator = (const Hyphenator &); + + linguistic::PropertyHelper_Hyphen & GetPropHelper_Impl(); + linguistic::PropertyHelper_Hyphen & GetPropHelper() + { + return pPropHelper ? *pPropHelper : GetPropHelper_Impl(); + } + +public: + Hyphenator(); + + virtual ~Hyphenator(); + + // XSupportedLocales (for XHyphenator) + virtual Sequence< Locale > SAL_CALL getLocales() + throw(RuntimeException); + virtual sal_Bool SAL_CALL hasLocale( const Locale& rLocale ) + throw(RuntimeException); + + // XHyphenator + virtual ::com::sun::star::uno::Reference< ::com::sun::star::linguistic2::XHyphenatedWord > SAL_CALL + hyphenate( const ::rtl::OUString& aWord, + const ::com::sun::star::lang::Locale& aLocale, + sal_Int16 nMaxLeading, + const ::com::sun::star::beans::PropertyValues& aProperties ) + throw(::com::sun::star::lang::IllegalArgumentException, + ::com::sun::star::uno::RuntimeException); + + virtual ::com::sun::star::uno::Reference< ::com::sun::star::linguistic2::XHyphenatedWord > SAL_CALL + queryAlternativeSpelling( const ::rtl::OUString& aWord, + const ::com::sun::star::lang::Locale& aLocale, + sal_Int16 nIndex, + const ::com::sun::star::beans::PropertyValues& aProperties ) + throw(::com::sun::star::lang::IllegalArgumentException, + ::com::sun::star::uno::RuntimeException); + + virtual ::com::sun::star::uno::Reference< ::com::sun::star::linguistic2::XPossibleHyphens > SAL_CALL + createPossibleHyphens( const ::rtl::OUString& aWord, + const ::com::sun::star::lang::Locale& aLocale, + const ::com::sun::star::beans::PropertyValues& aProperties ) + throw(::com::sun::star::lang::IllegalArgumentException, + ::com::sun::star::uno::RuntimeException); + + // XLinguServiceEventBroadcaster + virtual sal_Bool SAL_CALL + addLinguServiceEventListener( + const Reference< XLinguServiceEventListener >& rxLstnr ) + throw(RuntimeException); + virtual sal_Bool SAL_CALL + removeLinguServiceEventListener( + const Reference< XLinguServiceEventListener >& rxLstnr ) + throw(RuntimeException); + + // XServiceDisplayName + virtual OUString SAL_CALL + getServiceDisplayName( const Locale& rLocale ) + throw(RuntimeException); + + // XInitialization + virtual void SAL_CALL + initialize( const Sequence< Any >& rArguments ) + throw(Exception, RuntimeException); + + // XComponent + virtual void SAL_CALL + dispose() + throw(RuntimeException); + virtual void SAL_CALL + addEventListener( const Reference< XEventListener >& rxListener ) + throw(RuntimeException); + virtual void SAL_CALL + removeEventListener( const Reference< XEventListener >& rxListener ) + throw(RuntimeException); + + //////////////////////////////////////////////////////////// + // Service specific part + // + + // XServiceInfo + virtual OUString SAL_CALL + getImplementationName() + throw(RuntimeException); + virtual sal_Bool SAL_CALL + supportsService( const OUString& rServiceName ) + throw(RuntimeException); + virtual Sequence< OUString > SAL_CALL + getSupportedServiceNames() + throw(RuntimeException); + + + static inline OUString + getImplementationName_Static() throw(); + static Sequence< OUString > + getSupportedServiceNames_Static() throw(); + + + +private: + sal_uInt16 SAL_CALL capitalType(const OUString&, CharClass *); + OUString SAL_CALL makeLowerCase(const OUString&, CharClass *); + OUString SAL_CALL makeUpperCase(const OUString&, CharClass *); + OUString SAL_CALL makeInitCap(const OUString&, CharClass *); +}; + +inline OUString Hyphenator::getImplementationName_Static() throw() +{ + return A2OU( "org.openoffice.lingu.LibHnjHyphenator" ); +} + + +/////////////////////////////////////////////////////////////////////////// + +#endif + diff --git a/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/makefile.mk b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/makefile.mk new file mode 100644 index 000000000000..f3626c70312b --- /dev/null +++ b/lingucomponent/source/hyphenator/altlinuxhyph/hyphen/makefile.mk @@ -0,0 +1,93 @@ +#************************************************************************* +# +# DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. +# +# Copyright 2000, 2010 Oracle and/or its affiliates. +# +# OpenOffice.org - a multi-platform office productivity suite +# +# This file is part of OpenOffice.org. +# +# OpenOffice.org is free software: you can redistribute it and/or modify +# it under the terms of the GNU Lesser General Public License version 3 +# only, as published by the Free Software Foundation. +# +# OpenOffice.org is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU Lesser General Public License version 3 for more details +# (a copy is included in the LICENSE file that accompanied this code). +# +# You should have received a copy of the GNU Lesser General Public License +# version 3 along with OpenOffice.org. If not, see +# <http://www.openoffice.org/license.html> +# for a copy of the LGPLv3 License. +# +#************************************************************************* + +PRJ = ..$/..$/..$/.. + +PRJNAME = lingucomponent +TARGET = hyphen_lib +ENABLE_EXCEPTIONS=TRUE +USE_DEFFILE=TRUE + +.IF "$(GUI)"=="OS2" +HNJLIB=hyphen.lib +.ELIF "$(GUI)"=="UNX" || "$(COM)"=="GCC" +HNJLIB=-lhyphen +.ELSE +HNJLIB=hyphen.lib +.ENDIF + +#----- Settings --------------------------------------------------------- + +.INCLUDE : settings.mk + +# --- Files -------------------------------------------------------- + +.IF "$(SYSTEM_HUNSPELL)" != "YES" +HUNSPELL_CFLAGS += -I$(SOLARINCDIR)$/hunspell +.ENDIF + +CFLAGS += -I..$/..$/..$/lingutil $(HUNSPELL_CFLAGS) + +EXCEPTIONSFILES= \ + $(SLO)$/hyphenimp.obj + +SLOFILES= \ + $(SLO)$/hreg.obj\ + $(SLO)$/hyphenimp.obj + +REALNAME:=hyphen +SHL1TARGET= $(REALNAME)$(DLLPOSTFIX) + +SHL1STDLIBS= \ + $(CPPULIB) \ + $(CPPUHELPERLIB) \ + $(I18NISOLANGLIB) \ + $(TOOLSLIB) \ + $(SVLLIB) \ + $(SALLIB) \ + $(UNOTOOLSLIB) \ + $(LNGLIB) \ + $(HNJLIB) \ + $(ULINGULIB) \ + $(HUNSPELLLIB) + +# build DLL +SHL1LIBS= $(SLB)$/$(TARGET).lib $(SLB)$/libulingu.lib +SHL1IMPLIB= i$(REALNAME) +SHL1DEPN= $(SHL1LIBS) +SHL1DEF= $(MISC)$/$(SHL1TARGET).def + +SHL1VERSIONMAP=$(SOLARENV)/src/component.map + +# build DEF file +DEF1NAME =$(SHL1TARGET) +DEF1EXPORTFILE= exports.dxp + +# --- Targets ------------------------------------------------------ + +.INCLUDE : target.mk + |