Introduce SVBT16ToInt16 for reading sal_Int16 instead of sal_uInt16

This change does not change the actual results, but avoids Clang
-fsanitize=implicit-signed-integer-truncation warnings

Change-Id: If82193d7a6ad6bdadb4e8a8b3ed3bbdf3d88c4a3
Reviewed-on: https://gerrit.libreoffice.org/68706
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>

include/tools/solar.h

@@ -49,6 +49,7 @@ typedef sal_uInt8   SVBT64[8];
 inline sal_uInt16 SVBT16ToUInt16( const SVBT16 p ) { return static_cast<sal_uInt16>
                                                      (static_cast<sal_uInt16>(p[0])
                                                    + (static_cast<sal_uInt16>(p[1]) <<  8)); }
+inline sal_Int16 SVBT16ToInt16( const SVBT16 p ) { return sal_Int16(SVBT16ToUInt16(p)); }
 inline sal_uInt32 SVBT32ToUInt32 ( const SVBT32 p ) { return static_cast<sal_uInt32>
                                                      (static_cast<sal_uInt32>(p[0])
                                                    + (static_cast<sal_uInt32>(p[1]) <<  8)

sw/source/filter/ww8/ww8par6.cxx

@@ -144,7 +144,7 @@ static short ReadSprm( const WW8PLCFx_SEPX* pSep, sal_uInt16 nId, short nDefault
 {
     SprmResult aRes = pSep->HasSprm(nId);          // sprm here?
     const sal_uInt8* pS = aRes.pSprm;
-    short nVal = (pS && aRes.nRemainingData >= 2) ? SVBT16ToUInt16(pS) : nDefaultVal;
+    short nVal = (pS && aRes.nRemainingData >= 2) ? SVBT16ToInt16(pS) : nDefaultVal;
     return nVal;
 }
 
@@ -3439,7 +3439,7 @@ void SwWW8ImplReader::Read_SubSuperProp( sal_uInt16, const sal_uInt8* pData, sho
     }
 
     // font position in HalfPoints
-    short nPos = eVersion <= ww::eWW2 ? static_cast< sal_Int8 >( *pData ) : SVBT16ToUInt16( pData );
+    short nPos = eVersion <= ww::eWW2 ? static_cast< sal_Int8 >( *pData ) : SVBT16ToInt16( pData );
     sal_Int32 nPos2 = nPos * ( 10 * 100 );      // HalfPoints in 100 * tw
     const SvxFontHeightItem* pF
         = static_cast<const SvxFontHeightItem*>(GetFormatAttr(RES_CHRATR_FONTSIZE));