Resolves: fdo#87248 assume 0 chse means use encoding that matches lid

Change-Id: I813cc6bdc92b03bdfdd5202a85e958020a9c52c8

5 files changed, 25 insertions, 16 deletions

diff --git a/sw/source/filter/ww8/ww8glsy.cxx b/sw/source/filter/ww8/ww8glsy.cxx
index 2836518ece1c..cb60ebf19b13 100644
--- a/sw/source/filter/ww8/ww8glsy.cxx
+++ b/sw/source/filter/ww8/ww8glsy.cxx
@@ -197,7 +197,7 @@ bool WW8Glossary::Load( SwTextBlocks &rBlocks, bool bSaveRelFile )
         std::vector<ww::bytes> aData;
 
         rtl_TextEncoding eStructCharSet =
-            WW8Fib::GetFIBCharset(pGlossary->chseTables);
+            WW8Fib::GetFIBCharset(pGlossary->chseTables, pGlossary->lid);
 
         WW8ReadSTTBF(true, *xTableStream, pGlossary->fcSttbfglsy,
             pGlossary->lcbSttbfglsy, 0, eStructCharSet, aStrings, &aData );
diff --git a/sw/source/filter/ww8/ww8par.cxx b/sw/source/filter/ww8/ww8par.cxx
index 4d24c0aef61f..94962bc07ad3 100644
--- a/sw/source/filter/ww8/ww8par.cxx
+++ b/sw/source/filter/ww8/ww8par.cxx
@@ -5001,8 +5001,8 @@ sal_uLong SwWW8ImplReader::CoreLoad(WW8Glossary *pGloss, const SwPosition &rPos)
     bVer67 = bVer6 || bVer7;
     bVer8  = (8 == pWwFib->nVersion);
 
-    eTextCharSet = WW8Fib::GetFIBCharset(pWwFib->chse);
-    eStructCharSet = WW8Fib::GetFIBCharset(pWwFib->chseTables);
+    eTextCharSet = WW8Fib::GetFIBCharset(pWwFib->chse, pWwFib->lid);
+    eStructCharSet = WW8Fib::GetFIBCharset(pWwFib->chseTables, pWwFib->lid);
 
     bWWBugNormal = pWwFib->nProduct == 0xc03d;
 
@@ -5568,7 +5568,7 @@ namespace
             OUString sUniPassword = QueryPasswordForMedium( rMedium );
 
             OString sPassword(OUStringToOString(sUniPassword,
-                WW8Fib::GetFIBCharset(pWwFib->chseTables)));
+                WW8Fib::GetFIBCharset(pWwFib->chseTables, pWwFib->lid)));
 
             sal_Int32 nLen = sPassword.getLength();
             if( nLen <= 15 )
diff --git a/sw/source/filter/ww8/ww8par2.cxx b/sw/source/filter/ww8/ww8par2.cxx
index bbea4c013004..4929642615af 100644
--- a/sw/source/filter/ww8/ww8par2.cxx
+++ b/sw/source/filter/ww8/ww8par2.cxx
@@ -4216,7 +4216,7 @@ void WW8RStyle::ImportOldFormatStyles()
     }
 
     rtl_TextEncoding eStructChrSet = WW8Fib::GetFIBCharset(
-        pIo->pWwFib->chseTables);
+        pIo->pWwFib->chseTables, pIo->pWwFib->lid);
 
     sal_uInt16 cstcStd(0);
     rSt.ReadUInt16( cstcStd );
diff --git a/sw/source/filter/ww8/ww8scan.cxx b/sw/source/filter/ww8/ww8scan.cxx
index f5b71fbbc35b..dbc1de53304f 100644
--- a/sw/source/filter/ww8/ww8scan.cxx
+++ b/sw/source/filter/ww8/ww8scan.cxx
@@ -3984,7 +3984,7 @@ WW8PLCFx_Book::WW8PLCFx_Book(SvStream* pTblSt, const WW8Fib& rFib)
 
         pBook[1] = new WW8PLCFspecial(pTblSt,rFib.fcPlcfbkl,rFib.lcbPlcfbkl,0);
 
-        rtl_TextEncoding eStructChrSet = WW8Fib::GetFIBCharset(rFib.chseTables);
+        rtl_TextEncoding eStructChrSet = WW8Fib::GetFIBCharset(rFib.chseTables, rFib.lid);
 
         WW8ReadSTTBF( (7 < rFib.nVersion), *pTblSt, rFib.fcSttbfbkmk,
             rFib.lcbSttbfbkmk, 0, eStructChrSet, aBookNames );
@@ -5944,14 +5944,23 @@ bool WW8Fib::Write(SvStream& rStrm)
     return 0 == rStrm.GetError();
 }
 
-rtl_TextEncoding WW8Fib::GetFIBCharset(sal_uInt16 chs)
+rtl_TextEncoding WW8Fib::GetFIBCharset(sal_uInt16 chs, sal_uInt16 nLidLocale)
 {
     OSL_ENSURE(chs <= 0x100, "overflowed winword charset set");
-    rtl_TextEncoding eCharSet =
-        (0x0100 == chs)
-        ? RTL_TEXTENCODING_APPLE_ROMAN
-        : rtl_getTextEncodingFromWindowsCharset( static_cast<sal_uInt8>(chs) );
-    return eCharSet;
+    if (chs == 0x0100)
+        return RTL_TEXTENCODING_APPLE_ROMAN;
+    if (chs == 0 && nLidLocale != 0 && nLidLocale >= 999)
+    {
+        /*
+         nLidLocale:
+            language stamp -- localized version In pre-WinWord 2.0 files this
+            value was the nLocale. If value is < 999, then it is the nLocale,
+            otherwise it is the lid.
+        */
+        ::com::sun::star::lang::Locale aLocale(LanguageTag::convertToLocale(nLidLocale));
+        return msfilter::util::getBestTextEncodingFromLocale(aLocale);
+    }
+    return rtl_getTextEncodingFromWindowsCharset(static_cast<sal_uInt8>(chs));
 }
 
 WW8Style::WW8Style(SvStream& rStream, WW8Fib& rFibPara)
@@ -6334,7 +6343,7 @@ WW8Fonts::WW8Fonts( SvStream& rSt, WW8Fib& rFib )
                  the font, e.g load the doc in 97 and save to see the unicode
                  ver of the asian fontnames in that example to confirm.
                 */
-                rtl_TextEncoding eEnc = WW8Fib::GetFIBCharset(p->chs);
+                rtl_TextEncoding eEnc = WW8Fib::GetFIBCharset(p->chs, rFib.lid);
                 if ((eEnc == RTL_TEXTENCODING_SYMBOL) || (eEnc == RTL_TEXTENCODING_DONTKNOW))
                     eEnc = RTL_TEXTENCODING_MS_1252;
 
@@ -6364,7 +6373,7 @@ WW8Fonts::WW8Fonts( SvStream& rSt, WW8Fib& rFib )
                  the font, e.g load the doc in 97 and save to see the unicode
                  ver of the asian fontnames in that example to confirm.
                  */
-                rtl_TextEncoding eEnc = WW8Fib::GetFIBCharset(p->chs);
+                rtl_TextEncoding eEnc = WW8Fib::GetFIBCharset(p->chs, rFib.lid);
                 if ((eEnc == RTL_TEXTENCODING_SYMBOL) || (eEnc == RTL_TEXTENCODING_DONTKNOW))
                     eEnc = RTL_TEXTENCODING_MS_1252;
                 p->sFontname = OUString(pVer6->szFfn, rtl_str_getLength(pVer6->szFfn), eEnc);
@@ -6378,7 +6387,7 @@ WW8Fonts::WW8Fonts( SvStream& rSt, WW8Fib& rFib )
                 {
                     //#i18369# if it's a symbol font set Symbol as fallback
                     if (
-                         RTL_TEXTENCODING_SYMBOL == WW8Fib::GetFIBCharset(p->chs)
+                         RTL_TEXTENCODING_SYMBOL == WW8Fib::GetFIBCharset(p->chs, rFib.lid)
                          && p->sFontname!="Symbol"
                        )
                     {
diff --git a/sw/source/filter/ww8/ww8scan.hxx b/sw/source/filter/ww8/ww8scan.hxx
index 2dec722dbd64..dd3638b35714 100644
--- a/sw/source/filter/ww8/ww8scan.hxx
+++ b/sw/source/filter/ww8/ww8scan.hxx
@@ -1433,7 +1433,7 @@ public:
     WW8Fib( sal_uInt8 nVersion = 6, bool bDot = false );
     bool WriteHeader(SvStream& rStrm);
     bool Write(SvStream& rStrm);
-    static rtl_TextEncoding GetFIBCharset(sal_uInt16 chs);
+    static rtl_TextEncoding GetFIBCharset(sal_uInt16 chs, sal_uInt16 nLidLocale);
     ww::WordVersion GetFIBVersion() const;
     WW8_CP GetBaseCp(ManTypes nType) const;
     sal_Unicode getNumDecimalSep() const { return nNumDecimalSep;}