Resolves: fdo#66400 import combined characters from docx

move .doc combined character parser stuff from sw to filter for reuse in .docx

and fix bad length problem when nSavPtr == -1 after String->OUString conversion

thanks for the pasta CloudOn

Change-Id: I368ca30c14fd089271902b9b874de1099eb40038

2 files changed, 18 insertions, 290 deletions

diff --git a/sw/source/filter/ww8/ww8par.hxx b/sw/source/filter/ww8/ww8par.hxx
index 946a3d7c923c..7d970e6348ec 100644
--- a/sw/source/filter/ww8/ww8par.hxx
+++ b/sw/source/filter/ww8/ww8par.hxx
@@ -23,6 +23,7 @@
 #include <tools/string.hxx>
 #include "rtl/ustring.hxx"
 #include <filter/msfilter/msdffimp.hxx>
+#include <filter/msfilter/util.hxx>
 #include <editeng/frmdir.hxx>
 #include <fltshell.hxx>
 
@@ -100,7 +101,6 @@ class SwAttrSet;
 class GDIMetaFile;
 struct ESelection;
 class SfxItemSet;
-class WW8ReadFieldParams;
 class wwZOrderer;
 class OutlinerParaObject;
 
@@ -111,6 +111,8 @@ namespace com{namespace sun {namespace star{
     namespace lang{class XMultiServiceFactory;}
 }}}
 
+using namespace msfilter::util;
+
 // defines nur fuer die WW8-variable der INI-Datei
 #define WW8FL_NO_TEXT        1
 #define WW8FL_NO_STYLES      2
@@ -1771,7 +1773,6 @@ public:     // eigentlich private, geht aber leider nur public
     eF_ResT Read_F_DBNum( WW8FieldDesc*, OUString& );
     eF_ResT Read_F_Equation( WW8FieldDesc*, OUString& );
     void Read_SubF_Ruby( WW8ReadFieldParams& rReadParam);
-    void Read_SubF_Combined( WW8ReadFieldParams& rReadParam);
     eF_ResT Read_F_IncludePicture( WW8FieldDesc*, OUString& rStr );
     eF_ResT Read_F_IncludeText(    WW8FieldDesc*, OUString& rStr );
     eF_ResT Read_F_Seq( WW8FieldDesc*, OUString& rStr );
diff --git a/sw/source/filter/ww8/ww8par5.cxx b/sw/source/filter/ww8/ww8par5.cxx
index 885066683274..82ea7a4f8146 100644
--- a/sw/source/filter/ww8/ww8par5.cxx
+++ b/sw/source/filter/ww8/ww8par5.cxx
@@ -86,208 +86,12 @@
 #define WW8_TOX_LEVEL_DELIM     ':'
 
 using namespace ::com::sun::star;
+using namespace msfilter::util;
 using namespace sw::util;
 using namespace sw::mark;
 using namespace std; // #i24377#
 using namespace nsSwDocInfoSubType;
 
-
-class WW8ReadFieldParams
-{
-private:
-    const OUString aData;
-    sal_Int32 nFnd;
-    sal_Int32 nNext;
-    sal_Int32 nSavPtr;
-public:
-    WW8ReadFieldParams( const OUString& rData );
-    ~WW8ReadFieldParams();
-
-    bool GoToTokenParam();
-    sal_Int32 SkipToNextToken();
-    sal_Int32 GetTokenSttPtr() const   { return nFnd;  }
-
-    sal_Int32 FindNextStringPiece( sal_Int32 _nStart = -1 );
-    bool GetTokenSttFromTo(sal_Int32* _pFrom, sal_Int32* _pTo, sal_Int32 _nMax);
-
-    OUString GetResult() const;
-};
-
-
-WW8ReadFieldParams::WW8ReadFieldParams( const OUString& _rData )
-    : aData( _rData )
-    , nFnd( 0 )
-    , nNext( 0 )
-    , nSavPtr( 0 )
-{
-    /*
-        erstmal nach einer oeffnenden Klammer oder einer Leerstelle oder einem
-        Anfuehrungszeichen oder einem Backslash suchen, damit der Feldbefehl
-        (also INCLUDEPICTURE bzw EINFUeGENGRAFIK bzw ...) ueberlesen wird
-    */
-    const sal_Int32 nLen = aData.getLength();
-
-    while ( nNext<nLen && aData[nNext]==' ' )
-        ++nNext;
-
-    while ( nNext<nLen )
-    {
-        const sal_Unicode c = aData[nNext];
-        if ( c==' ' || c=='"' || c=='\\' || c==132 || c==0x201c )
-            break;
-        ++nNext;
-    }
-
-    nFnd      = nNext;
-    nSavPtr   = nNext;
-}
-
-
-WW8ReadFieldParams::~WW8ReadFieldParams()
-{
-
-}
-
-
-OUString WW8ReadFieldParams::GetResult() const
-{
-    return nFnd<0 && nSavPtr>nFnd ? OUString() : aData.copy( nFnd, nSavPtr-nFnd );
-}
-
-
-bool WW8ReadFieldParams::GoToTokenParam()
-{
-    const sal_Int32 nOld = nNext;
-    if( -2 == SkipToNextToken() )
-        return GetTokenSttPtr()>=0;
-    nNext = nOld;
-    return false;
-}
-
-// ret: -2: NOT a '\' parameter but normal Text
-sal_Int32 WW8ReadFieldParams::SkipToNextToken()
-{
-    if ( nNext<0 || nNext>=aData.getLength() )
-        return -1;
-
-    nFnd = FindNextStringPiece(nNext);
-    if ( nFnd<0 )
-        return -1;
-
-    nSavPtr = nNext;
-
-    if ( aData[nFnd]=='\\' && nFnd+1<aData.getLength() && aData[nFnd+1]!='\\' )
-    {
-        const sal_Int32 nRet = aData[++nFnd];
-        nNext = ++nFnd;             // und dahinter setzen
-        return nRet;
-    }
-
-    if ( nSavPtr>0 && (aData[nSavPtr-1]=='"' || aData[nSavPtr-1]==0x201d ) )
-    {
-        --nSavPtr;
-    }
-    return -2;
-}
-
-// FindNextPara sucht naechsten Backslash-Parameter oder naechste Zeichenkette
-// bis zum Blank oder naechsten "\" oder zum schliessenden Anfuehrungszeichen
-// oder zum String-Ende von pStr.
-//
-// Ausgabe ppNext (falls ppNext != 0) Suchbeginn fuer naechsten Parameter bzw. 0
-//
-// Returnwert: 0 falls String-Ende erreicht,
-//             ansonsten Anfang des Paramters bzw. der Zeichenkette
-//
-sal_Int32 WW8ReadFieldParams::FindNextStringPiece(const sal_Int32 nStart)
-{
-    const sal_Int32 nLen = aData.getLength();
-    sal_Int32  n = nStart<0  ? nFnd : nStart;  // Anfang
-    sal_Int32 n2;          // Ende
-
-    nNext = -1;        // Default fuer nicht gefunden
-
-    while ( n<nLen && aData[n]==' ' )
-        ++n;
-
-    if ( n==nLen )
-        return -1;
-
-    if ( aData[n]==0x13 )
-    {
-        // Skip the nested field code since it's not supported
-        while ( n<nLen && aData[n]!=0x14 )
-            ++n;
-        if ( n==nLen )
-            return -1;
-    }
-
-    // Anfuehrungszeichen vor Para?
-    if ( aData[n]=='"' || aData[n]==0x201c || aData[n]==132 || aData[n]==0x14 )
-    {
-        n++;                        // Anfuehrungszeichen ueberlesen
-        n2 = n;                     // ab hier nach Ende suchen
-        while(     (nLen > n2)
-                && (aData[n2] != '"')
-                && (aData[n2] != 0x201d)
-                && (aData[n2] != 147)
-                && (aData[n2] != 0x15) )
-            n2++;                   // Ende d. Paras suchen
-    }
-    else                        // keine Anfuehrungszeichen
-    {
-        n2 = n;                     // ab hier nach Ende suchen
-        while ( n2<nLen && aData[n2]!=' ' ) // Ende d. Paras suchen
-        {
-            if ( aData[n2]=='\\' )
-            {
-                if ( n2+1<nLen && aData[n2+1]=='\\' )
-                    n2 += 2;        // Doppel-Backslash -> OK
-                else
-                {
-                    if( n2 > n )
-                        n2--;
-                    break;          // einfach-Backslash -> Ende
-                }
-            }
-            else
-                n2++;               // kein Backslash -> OK
-        }
-    }
-    if( nLen > n2 )
-    {
-        if (aData[n2]!=' ') ++n2;
-        nNext = n2;
-    }
-    return n;
-}
-
-
-
-// read parameters "1-3" or 1-3 with both values between 1 and nMax
-bool WW8ReadFieldParams::GetTokenSttFromTo(sal_Int32* pFrom, sal_Int32* pTo, sal_Int32 nMax)
-{
-    sal_Int32 nStart = 0;
-    sal_Int32 nEnd   = 0;
-    if ( GoToTokenParam() )
-    {
-
-        const OUString sParams( GetResult() );
-
-        sal_Int32 nIndex = 0;
-        const OUString sStart( sParams.getToken(0, '-', nIndex) );
-        if (nIndex>=0)
-        {
-            nStart = sStart.toInt32();
-            nEnd   = sParams.copy(nIndex).toInt32();
-        }
-    }
-    if( pFrom ) *pFrom = nStart;
-    if( pTo )   *pTo   = nEnd;
-
-    return nStart && nEnd && (nMax >= nStart) && (nMax >= nEnd);
-}
-
 //----------------------------------------
 //              Bookmarks
 //----------------------------------------
@@ -2608,103 +2412,26 @@ eF_ResT SwWW8ImplReader::Read_F_Equation( WW8FieldDesc*, OUString& rStr )
     WW8ReadFieldParams aReadParam( rStr );
     const sal_Int32 cChar = aReadParam.SkipToNextToken();
     if ('o' == cChar || 'O' == cChar)
-        Read_SubF_Combined(aReadParam);
-    else if ('*' == cChar)
-        Read_SubF_Ruby(aReadParam);
-    return FLD_OK;
-}
-
-void SwWW8ImplReader::Read_SubF_Combined( WW8ReadFieldParams& rReadParam)
-{
-    String sCombinedCharacters;
-    WW8ReadFieldParams aOriFldParam = rReadParam;
-    const sal_Int32 cGetChar = rReadParam.SkipToNextToken();
-    switch( cGetChar )
     {
-    case 'a':
-    case 'A':
-        if ( !rReadParam.GetResult().startsWithIgnoreAsciiCase("d") )
+        EquationResult aResult(ParseCombinedChars(rStr));
+
+        if (aResult.sType == "Input")
         {
-            break;
+            SwInputField aFld( (SwInputFieldType*)rDoc.GetSysFldType( RES_INPUTFLD ),
+                aResult.sResult, aResult.sResult, INP_TXT, 0 );
+            rDoc.InsertPoolItem( *pPaM, SwFmtFld( aFld ), 0 ); // insert input field
         }
-        rReadParam.SkipToNextToken();
-        // intentional fall-through
-    case -2:
+        else if (aResult.sType == "CombinedCharacters")
         {
-            if ( rReadParam.GetResult().startsWithIgnoreAsciiCase("(") )
-            {
-                for (int i=0;i<2;i++)
-                {
-                    if ('s' == rReadParam.SkipToNextToken())
-                    {
-                        const sal_Int32 cChar = rReadParam.SkipToNextToken();
-                        if (-2 != rReadParam.SkipToNextToken())
-                            break;
-                        const OUString sF = rReadParam.GetResult();
-                        if ((('u' == cChar) && sF.startsWithIgnoreAsciiCase("p"))
-                            || (('d' == cChar) && sF.startsWithIgnoreAsciiCase("o")))
-                        {
-                            if (-2 == rReadParam.SkipToNextToken())
-                            {
-                                String sPart = rReadParam.GetResult();
-                                xub_StrLen nBegin = sPart.Search('(');
-
-                                //Word disallows brackets in this field, which
-                                //aids figuring out the case of an end of )) vs )
-                                xub_StrLen nEnd = sPart.Search(')');
-
-                                if ((nBegin != STRING_NOTFOUND) &&
-                                    (nEnd != STRING_NOTFOUND))
-                                {
-                                    sCombinedCharacters +=
-                                        sPart.Copy(nBegin+1,nEnd-nBegin-1);
-                                }
-                            }
-                        }
-                    }
-                }
-                if (sCombinedCharacters.Len())
-                {
-                    SwCombinedCharField aFld((SwCombinedCharFieldType*)
-                        rDoc.GetSysFldType(RES_COMBINED_CHARS),sCombinedCharacters);
-                    rDoc.InsertPoolItem(*pPaM, SwFmtFld(aFld), 0);
-                }
-                else
-                {
-                    const String sPart = aOriFldParam.GetResult();
-                    xub_StrLen nBegin = sPart.Search('(');
-                    xub_StrLen nEnd = sPart.Search(',');
-                    if ( nEnd == STRING_NOTFOUND )
-                    {
-                        nEnd = sPart.Search(')');
-                    }
-                    if ( (nBegin != STRING_NOTFOUND) && (nEnd != STRING_NOTFOUND) )
-                    {
-                        // skip certain leading characters
-                        for (int i = nBegin;i < nEnd-1;i++)
-                        {
-                            const sal_Unicode cC = sPart.GetChar(nBegin+1);
-                            if ( cC < 32 )
-                            {
-                                nBegin++;
-                            }
-                            else
-                                break;
-                        }
-                        sCombinedCharacters = sPart.Copy( nBegin+1, nEnd-nBegin-1 );
-                        if ( sCombinedCharacters.Len() )
-                        {
-                            SwInputField aFld( (SwInputFieldType*)rDoc.GetSysFldType( RES_INPUTFLD ),
-                                sCombinedCharacters, sCombinedCharacters, INP_TXT, 0 );
-                            rDoc.InsertPoolItem( *pPaM, SwFmtFld( aFld ), 0 ); // insert input field
-                        }
-                    }
-                }
-            }
+            SwCombinedCharField aFld((SwCombinedCharFieldType*)
+                rDoc.GetSysFldType(RES_COMBINED_CHARS), aResult.sType);
+            rDoc.InsertPoolItem(*pPaM, SwFmtFld(aFld), 0);
         }
-    default:
-        break;
     }
+    else if ('*' == cChar)
+        Read_SubF_Ruby(aReadParam);
+
+    return FLD_OK;
 }
 
 void SwWW8ImplReader::Read_SubF_Ruby( WW8ReadFieldParams& rReadParam)