diff options
Diffstat (limited to 'sw/source/core/txtnode/txtedt.cxx')
-rw-r--r-- | sw/source/core/txtnode/txtedt.cxx | 271 |
1 files changed, 238 insertions, 33 deletions
diff --git a/sw/source/core/txtnode/txtedt.cxx b/sw/source/core/txtnode/txtedt.cxx index 944eb70b2c3d..0630dbd722b6 100644 --- a/sw/source/core/txtnode/txtedt.cxx +++ b/sw/source/core/txtnode/txtedt.cxx @@ -36,6 +36,7 @@ #include <time.h> // clock() #include <tools/stream.hxx> #endif + #include <hintids.hxx> #include <vcl/svapp.hxx> #include <svl/itemiter.hxx> @@ -46,9 +47,6 @@ #include <editeng/hangulhanja.hxx> #include <SwSmartTagMgr.hxx> #include <linguistic/lngprops.hxx> -#include <com/sun/star/beans/XPropertySet.hpp> -#include <com/sun/star/i18n/WordType.hdl> -#include <com/sun/star/i18n/ScriptType.hdl> #include <unotools/transliterationwrapper.hxx> #include <unotools/charclass.hxx> #include <dlelstnr.hxx> @@ -86,6 +84,15 @@ #include <unomid.h> +#include <com/sun/star/beans/XPropertySet.hpp> +#include <com/sun/star/i18n/WordType.hdl> +#include <com/sun/star/i18n/ScriptType.hdl> +#include <com/sun/star/i18n/TransliterationModules.hpp> +#include <com/sun/star/i18n/TransliterationModulesExtra.hpp> + +#include <vector> + + using rtl::OUString; using namespace ::com::sun::star; using namespace ::com::sun::star::frame; @@ -1578,49 +1585,247 @@ void SwLinguStatistik::Flush() #endif + +struct TransliterationChgData +{ + xub_StrLen nStart; + xub_StrLen nLen; + String sChanged; + Sequence< sal_Int32 > aOffsets; +}; + // change text to Upper/Lower/Hiragana/Katagana/... -void SwTxtNode::TransliterateText( utl::TransliterationWrapper& rTrans, - xub_StrLen nStt, xub_StrLen nEnd, SwUndoTransliterate* pUndo ) +void SwTxtNode::TransliterateText( + utl::TransliterationWrapper& rTrans, + xub_StrLen nStt, xub_StrLen nEnd, + SwUndoTransliterate* pUndo ) { - if( nStt < nEnd ) + if (nStt < nEnd && pBreakIt->GetBreakIter().is()) { - SwLanguageIterator* pIter; - if( rTrans.needLanguageForTheMode() ) - pIter = new SwLanguageIterator( *this, nStt ); - else - pIter = 0; + // since we don't use Hiragana/Katakana or half-width/full-width transliterations here + // it is fine to use ANYWORD_IGNOREWHITESPACES. (ANY_WORD btw is broken and will + // occasionaly miss words in consecutive sentences). Also with ANYWORD_IGNOREWHITESPACES + // text like 'just-in-time' will be converted to 'Just-In-Time' which seems to be the + // proper thing to do. + const sal_Int16 nWordType = WordType::ANYWORD_IGNOREWHITESPACES; + + //! In order to have less trouble with changing text size, e.g. because + //! of ligatures or � (German small sz) being resolved, we need to process + //! the text replacements from end to start. + //! This way the offsets for the yet to be changed words will be + //! left unchanged by the already replaced text. + //! For this we temporarily save the changes to be done in this vector + std::vector< TransliterationChgData > aChanges; + TransliterationChgData aChgData; + + if (rTrans.getType() == (sal_uInt32)TransliterationModulesExtra::TITLE_CASE) + { + // for 'capitalize every word' we need to iterate over each word + + Boundary aSttBndry; + Boundary aEndBndry; + aSttBndry = pBreakIt->GetBreakIter()->getWordBoundary( + GetTxt(), nStt, + pBreakIt->GetLocale( GetLang( nStt ) ), + nWordType, + TRUE /*prefer forward direction*/); + aEndBndry = pBreakIt->GetBreakIter()->getWordBoundary( + GetTxt(), nEnd, + pBreakIt->GetLocale( GetLang( nEnd ) ), + nWordType, + FALSE /*prefer backward direction*/); + + // prevent backtracking to the previous word if selection is at word boundary + if (aSttBndry.endPos <= nStt) + { + aSttBndry = pBreakIt->GetBreakIter()->nextWord( + GetTxt(), aSttBndry.endPos, + pBreakIt->GetLocale( GetLang( aSttBndry.endPos ) ), + nWordType); + } + // prevent advancing to the next word if selection is at word boundary + if (aEndBndry.startPos >= nEnd) + { + aEndBndry = pBreakIt->GetBreakIter()->previousWord( + GetTxt(), aEndBndry.startPos, + pBreakIt->GetLocale( GetLang( aEndBndry.startPos ) ), + nWordType); + } - xub_StrLen nEndPos; - sal_uInt16 nLang; - do { - if( pIter ) + Boundary aCurWordBndry( aSttBndry ); + while (aCurWordBndry.startPos <= aEndBndry.startPos) { - nLang = pIter->GetLanguage(); - nEndPos = pIter->GetChgPos(); - if( nEndPos > nEnd ) - nEndPos = nEnd; + nStt = (xub_StrLen)aCurWordBndry.startPos; + nEnd = (xub_StrLen)aCurWordBndry.endPos; + sal_Int32 nLen = nEnd - nStt; + DBG_ASSERT( nLen > 0, "invalid word length of 0" ); +#if OSL_DEBUG_LEVEL > 1 + String aText( GetTxt().Copy( nStt, nLen ) ); +#endif + + Sequence <sal_Int32> aOffsets; + String sChgd( rTrans.transliterate( GetTxt(), GetLang( nStt ), nStt, nLen, &aOffsets )); + + if (!m_Text.Equals( sChgd, nStt, nLen )) + { + aChgData.nStart = nStt; + aChgData.nLen = nLen; + aChgData.sChanged = sChgd; + aChgData.aOffsets = aOffsets; + aChanges.push_back( aChgData ); + } + + aCurWordBndry = pBreakIt->GetBreakIter()->nextWord( + GetTxt(), nEnd, + pBreakIt->GetLocale( GetLang( nEnd ) ), + nWordType); } - else + } + else if (rTrans.getType() == (sal_uInt32)TransliterationModulesExtra::SENTENCE_CASE) + { + // for 'sentence case' we need to iterate sentence by sentence + + sal_Int32 nLastStart = pBreakIt->GetBreakIter()->beginOfSentence( + GetTxt(), nEnd, + pBreakIt->GetLocale( GetLang( nEnd ) ) ); + sal_Int32 nLastEnd = pBreakIt->GetBreakIter()->endOfSentence( + GetTxt(), nLastStart, + pBreakIt->GetLocale( GetLang( nLastStart ) ) ); + + // extend nStt, nEnd to the current sentence boundaries + sal_Int32 nCurrentStart = pBreakIt->GetBreakIter()->beginOfSentence( + GetTxt(), nStt, + pBreakIt->GetLocale( GetLang( nStt ) ) ); + sal_Int32 nCurrentEnd = pBreakIt->GetBreakIter()->endOfSentence( + GetTxt(), nCurrentStart, + pBreakIt->GetLocale( GetLang( nCurrentStart ) ) ); + + // prevent backtracking to the previous sentence if selection starts at end of a sentence + if (nCurrentEnd <= nStt) + { + // now nCurrentStart is probably located on a non-letter word. (unless we + // are in Asian text with no spaces...) + // Thus to get the real sentence start we should locate the next real word, + // that is one found by DICTIONARY_WORD + i18n::Boundary aBndry = pBreakIt->GetBreakIter()->nextWord( + GetTxt(), nCurrentEnd, + pBreakIt->GetLocale( GetLang( nCurrentEnd ) ), + i18n::WordType::DICTIONARY_WORD); + + // now get new current sentence boundaries + nCurrentStart = pBreakIt->GetBreakIter()->beginOfSentence( + GetTxt(), aBndry.startPos, + pBreakIt->GetLocale( GetLang( aBndry.startPos) ) ); + nCurrentEnd = pBreakIt->GetBreakIter()->endOfSentence( + GetTxt(), nCurrentStart, + pBreakIt->GetLocale( GetLang( nCurrentStart) ) ); + } + // prevent advancing to the next sentence if selection ends at start of a sentence + if (nLastStart >= nEnd) { - nLang = LANGUAGE_SYSTEM; - nEndPos = nEnd; + // now nCurrentStart is probably located on a non-letter word. (unless we + // are in Asian text with no spaces...) + // Thus to get the real sentence start we should locate the previous real word, + // that is one found by DICTIONARY_WORD + i18n::Boundary aBndry = pBreakIt->GetBreakIter()->previousWord( + GetTxt(), nLastStart, + pBreakIt->GetLocale( GetLang( nLastStart) ), + i18n::WordType::DICTIONARY_WORD); + nLastEnd = pBreakIt->GetBreakIter()->endOfSentence( + GetTxt(), aBndry.startPos, + pBreakIt->GetLocale( GetLang( aBndry.startPos) ) ); + if (nCurrentEnd > nLastEnd) + nCurrentEnd = nLastEnd; } - xub_StrLen nLen = nEndPos - nStt; - Sequence <sal_Int32> aOffsets; - String sChgd( rTrans.transliterate( m_Text, nLang, nStt, nLen, - &aOffsets )); - if( !m_Text.Equals( sChgd, nStt, nLen ) ) + while (nCurrentStart < nLastEnd) { - if ( pUndo ) + sal_Int32 nLen = nCurrentEnd - nCurrentStart; + DBG_ASSERT( nLen > 0, "invalid word length of 0" ); +#if OSL_DEBUG_LEVEL > 1 + String aText( GetTxt().Copy( nCurrentStart, nLen ) ); +#endif + + Sequence <sal_Int32> aOffsets; + String sChgd( rTrans.transliterate( GetTxt(), + GetLang( nCurrentStart ), nCurrentStart, nLen, &aOffsets )); + + if (!m_Text.Equals( sChgd, nStt, nLen )) { - pUndo->AddChanges( *this, nStt, nLen, aOffsets ); + aChgData.nStart = nCurrentStart; + aChgData.nLen = nLen; + aChgData.sChanged = sChgd; + aChgData.aOffsets = aOffsets; + aChanges.push_back( aChgData ); } - ReplaceTextOnly( nStt, nLen, sChgd, aOffsets ); + + Boundary aFirstWordBndry; + aFirstWordBndry = pBreakIt->GetBreakIter()->nextWord( + GetTxt(), nCurrentEnd, + pBreakIt->GetLocale( GetLang( nCurrentEnd ) ), + nWordType); + nCurrentStart = aFirstWordBndry.startPos; + nCurrentEnd = pBreakIt->GetBreakIter()->endOfSentence( + GetTxt(), nCurrentStart, + pBreakIt->GetLocale( GetLang( nCurrentStart ) ) ); } - nStt = nEndPos; - } while( nEndPos < nEnd && pIter && pIter->Next() ); - delete pIter; + } + else + { + // here we may transliterate over complete language portions... + + SwLanguageIterator* pIter; + if( rTrans.needLanguageForTheMode() ) + pIter = new SwLanguageIterator( *this, nStt ); + else + pIter = 0; + + xub_StrLen nEndPos; + sal_uInt16 nLang; + do { + if( pIter ) + { + nLang = pIter->GetLanguage(); + nEndPos = pIter->GetChgPos(); + if( nEndPos > nEnd ) + nEndPos = nEnd; + } + else + { + nLang = LANGUAGE_SYSTEM; + nEndPos = nEnd; + } + xub_StrLen nLen = nEndPos - nStt; + + Sequence <sal_Int32> aOffsets; + String sChgd( rTrans.transliterate( m_Text, nLang, nStt, nLen, &aOffsets )); + + if (!m_Text.Equals( sChgd, nStt, nLen )) + { + aChgData.nStart = nStt; + aChgData.nLen = nLen; + aChgData.sChanged = sChgd; + aChgData.aOffsets = aOffsets; + aChanges.push_back( aChgData ); + } + + nStt = nEndPos; + } while( nEndPos < nEnd && pIter && pIter->Next() ); + delete pIter; + } + + if (aChanges.size() > 0) + { + // now apply the changes from end to start to leave the offsets of the + // yet unchanged text parts remain the same. + for (size_t i = 0; i < aChanges.size(); ++i) + { + TransliterationChgData &rData = aChanges[ aChanges.size() - 1 - i ]; + if (pUndo) + pUndo->AddChanges( *this, rData.nStart, rData.nLen, rData.aOffsets ); + ReplaceTextOnly( rData.nStart, rData.nLen, rData.sChanged, rData.aOffsets ); + } + } } } |