tdf#152425 Make Word names unique, and use them for style ids generation

Before, a style id was generated from LibreOffice name, and then the
name was replaced with Word name. Given that Writer's List N maps to
Word's List Bullet N, and Word's List N doesn't map to anything in
Writer, this led to styles.xml having after roundtrip:

<w:style w:type="paragraph" w:styleId="List5">
<w:name w:val="List Bullet 5"/>
...
<w:style w:type="paragraph" w:styleId="ListBullet5">
<w:name w:val="List Bullet 5"/>

So the idea is to do the following steps:

1. Collect all the exported styles (unchanged);
2. Build unique Word names for collected styles (new):
a. Process all the styles that map to special Word styles first,
so that their Word names don't get changed when made unique;
b. Process the rest of the styles, making sure to append a sequential
number after the Writer name, if a clash happens.
3. Build Style Ids from the Word names (previously Writer name could be
used), also making sure they are unique.

Change-Id: I9f8f254aa6ae713671234f0109b94cc72a588150
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/143905
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/145376
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>

6 files changed, 101 insertions, 82 deletions

diff --git a/sw/qa/extras/ooxmlexport/ooxmlexport13.cxx b/sw/qa/extras/ooxmlexport/ooxmlexport13.cxx
index 671d9c9958db..cd194b5c68f1 100644
--- a/sw/qa/extras/ooxmlexport/ooxmlexport13.cxx
+++ b/sw/qa/extras/ooxmlexport/ooxmlexport13.cxx
@@ -1238,8 +1238,8 @@ CPPUNIT_TEST_FIXTURE(Test, testTdf123628)
 
     xmlDocUniquePtr pXmlStyles = parseExport("word/styles.xml");
 
-    assertXPath(pXmlDoc, "/w:document/w:body/w:p[1]/w:hyperlink/w:r/w:rPr/w:rStyle", "val", "InternetLink");
-    assertXPath(pXmlStyles, "/w:styles/w:style[@w:styleId='InternetLink']/w:name", "val", "Hyperlink");
+    assertXPath(pXmlDoc, "/w:document/w:body/w:p[1]/w:hyperlink/w:r/w:rPr/w:rStyle", "val", "Hyperlink");
+    assertXPath(pXmlStyles, "/w:styles/w:style[@w:styleId='Hyperlink']/w:name", "val", "Hyperlink");
 }
 
 DECLARE_OOXMLEXPORT_TEST(testTdf127741, "tdf127741.docx")
@@ -1265,7 +1265,7 @@ CPPUNIT_TEST_FIXTURE(Test, testTdf127925)
     loadAndSave("tdf127925.odt");
     CPPUNIT_ASSERT_EQUAL(1, getPages());
     xmlDocUniquePtr pXmlStyles = parseExport("word/styles.xml");
-    assertXPath(pXmlStyles, "/w:styles/w:style[@w:styleId='VisitedInternetLink']/w:name", "val", "FollowedHyperlink");
+    assertXPath(pXmlStyles, "/w:styles/w:style[@w:styleId='FollowedHyperlink']/w:name", "val", "FollowedHyperlink");
 }
 
 CPPUNIT_TEST_FIXTURE(Test, testTdf127579)
@@ -1273,7 +1273,7 @@ CPPUNIT_TEST_FIXTURE(Test, testTdf127579)
     loadAndSave("tdf127579.odt");
     CPPUNIT_ASSERT_EQUAL(1, getPages());
     xmlDocUniquePtr pXmlDoc = parseExport("word/document.xml");
-    assertXPath(pXmlDoc, "/w:document/w:body/w:p/w:hyperlink/w:r/w:rPr/w:rStyle", "val", "InternetLink");
+    assertXPath(pXmlDoc, "/w:document/w:body/w:p/w:hyperlink/w:r/w:rPr/w:rStyle", "val", "Hyperlink");
 }
 
 DECLARE_OOXMLEXPORT_TEST(testTdf128304, "tdf128304.odt")
diff --git a/sw/qa/extras/ooxmlexport/ooxmlexport16.cxx b/sw/qa/extras/ooxmlexport/ooxmlexport16.cxx
index ae8b96c99f66..22c3b75c5e87 100644
--- a/sw/qa/extras/ooxmlexport/ooxmlexport16.cxx
+++ b/sw/qa/extras/ooxmlexport/ooxmlexport16.cxx
@@ -984,7 +984,7 @@ DECLARE_OOXMLEXPORT_TEST(testTdf143726, "Simple-TOC.odt")
     CPPUNIT_ASSERT(pXmlStyles);
     // Without the fix this was "TOA Heading" which belongs to the "Table of Authorities" index in Word
     // TOC's heading style should be exported as "TOC Heading" as that's the default Word style name
-    assertXPath(pXmlStyles, "/w:styles/w:style[@w:styleId='ContentsHeading']/w:name", "val", "TOC Heading");
+    assertXPath(pXmlStyles, "/w:styles/w:style[@w:styleId='TOCHeading']/w:name", "val", "TOC Heading");
 }
 
 CPPUNIT_PLUGIN_IMPLEMENT();
diff --git a/sw/qa/extras/ooxmlexport/ooxmlexport18.cxx b/sw/qa/extras/ooxmlexport/ooxmlexport18.cxx
index 37b1bad5d10e..97d1d117e3ea 100644
--- a/sw/qa/extras/ooxmlexport/ooxmlexport18.cxx
+++ b/sw/qa/extras/ooxmlexport/ooxmlexport18.cxx
@@ -153,6 +153,11 @@ CPPUNIT_TEST_FIXTURE(Test, testTdf152425)
     // CPPUNIT_ASSERT_EQUAL(OUString("List 5"), Para4Style);
     // But for now, just make sure that the style names differ
     CPPUNIT_ASSERT(Para4Style != Para3Style);
+    const OUString Para5Style = getProperty<OUString>(getParagraph(5), "ParaStyleName");
+    // Eventually, we need to check this:
+    // CPPUNIT_ASSERT_EQUAL(OUString("List Bullet 5"), Para5Style);
+    // But for now, just make sure that the style names differ
+    CPPUNIT_ASSERT(Para5Style != Para4Style);
 }
 
 CPPUNIT_PLUGIN_IMPLEMENT();
diff --git a/sw/source/filter/ww8/docxattributeoutput.cxx b/sw/source/filter/ww8/docxattributeoutput.cxx
index b6334c48af83..e3a7828e5961 100644
--- a/sw/source/filter/ww8/docxattributeoutput.cxx
+++ b/sw/source/filter/ww8/docxattributeoutput.cxx
@@ -3861,7 +3861,12 @@ void DocxAttributeOutput::Redline( const SwRedlineData* pRedlineData)
 
                     m_pSerializer->startElementNS(XML_w, XML_pPr);
 
-                    OString sStyleName = MSWordStyles::CreateStyleId( sParaStyleName );
+                    OString sStyleName;
+                    if (auto format = m_rExport.m_rDoc.FindTextFormatCollByName(sParaStyleName))
+                        if (auto slot = m_rExport.m_pStyles->GetSlot(format); slot != 0xfff)
+                            sStyleName = m_rExport.m_pStyles->GetStyleId(slot);
+                    if (sStyleName.isEmpty())
+                        sStyleName = MSWordStyles::CreateStyleId(sParaStyleName);
                     if ( !sStyleName.isEmpty() )
                         m_pSerializer->singleElementNS(XML_w, XML_pStyle, FSNS(XML_w, XML_val), sStyleName);
 
@@ -7204,20 +7209,14 @@ void DocxAttributeOutput::StartStyle( const OUString& rName, StyleType eType,
             SAL_WARN("sw.ww8", "Unhandled style property: " << rProp.Name);
     }
 
-    // MSO exports English names and writerfilter only recognize them.
-    const char *pEnglishName = nullptr;
     const char* pType = nullptr;
     switch (eType)
     {
         case STYLE_TYPE_PARA:
             pType = "paragraph";
-            if ( nWwId < ww::stiMax)
-                pEnglishName = ww::GetEnglishNameFromSti( static_cast<ww::sti>(nWwId ) );
             break;
         case STYLE_TYPE_CHAR:
             pType = "character";
-            if (nWwId < ww::stiMax)
-                pEnglishName = ww::GetEnglishNameFromSti(static_cast<ww::sti>(nWwId));
             break;
         case STYLE_TYPE_LIST: pType = "numbering"; break;
     }
@@ -7228,8 +7227,7 @@ void DocxAttributeOutput::StartStyle( const OUString& rName, StyleType eType,
     if (bCustomStyle)
         pStyleAttributeList->add(FSNS(XML_w, XML_customStyle), "1");
     m_pSerializer->startElementNS( XML_w, XML_style, pStyleAttributeList);
-    m_pSerializer->singleElementNS( XML_w, XML_name,
-            FSNS( XML_w, XML_val ), pEnglishName ? pEnglishName : rName.toUtf8() );
+    m_pSerializer->singleElementNS(XML_w, XML_name, FSNS(XML_w, XML_val), rName);
 
     if ( nBase != 0x0FFF && eType != STYLE_TYPE_LIST)
     {
diff --git a/sw/source/filter/ww8/wrtw8sty.cxx b/sw/source/filter/ww8/wrtw8sty.cxx
index 98730c514718..b76907229f0d 100644
--- a/sw/source/filter/ww8/wrtw8sty.cxx
+++ b/sw/source/filter/ww8/wrtw8sty.cxx
@@ -156,6 +156,7 @@ MSWordStyles::MSWordStyles( MSWordExportBase& rExport, bool bListStyles )
     memset( m_aHeadingParagraphStyles, -1 , MAXLEVEL * sizeof( sal_uInt16));
 
     BuildStylesTable();
+    BuildWwNames();
     BuildStyleIds();
 }
 
@@ -286,7 +287,7 @@ void MSWordStyles::BuildStylesTable()
         sal_uInt16 nSlot = BuildGetSlot(*pFormat);
         if (nSlot != 0xfff)
         {
-            m_aStyles[nSlot].format = pFormat;
+            m_aStyles[nSlot] = { pFormat };
         }
         else
         {
@@ -316,7 +317,60 @@ void MSWordStyles::BuildStylesTable()
     }
 }
 
-OString MSWordStyles::CreateStyleId(const OUString &rName)
+void MSWordStyles::BuildWwNames()
+{
+    std::unordered_set<OUString> aUsed;
+
+    auto makeUniqueName = [&aUsed](OUString& name) {
+        // toAsciiLowerCase rules out e.g. user's "normal"; no problem if there are non-ASCII chars
+        OUString lower(name.toAsciiLowerCase());
+        if (!aUsed.insert(lower).second)
+        {
+            int nFree = 1;
+            while (!aUsed.insert(lower + OUString::number(nFree)).second)
+                ++nFree;
+
+            name += OUString::number(nFree);
+        }
+    };
+
+    // We want to map LO's default style to Word's "Normal" style.
+    // Word looks for this specific style name when reading docx files.
+    // (It must be the English word regardless of language settings)
+    assert(!m_aStyles.empty());
+    assert(!m_aStyles[0].format || m_aStyles[0].ww_id == ww::stiNormal);
+    m_aStyles[0].ww_name = "Normal";
+    aUsed.insert("normal");
+
+    // 1. Handle styles having special wwIds, and thus pre-defined names
+    for (auto& entry : m_aStyles)
+    {
+        if (!entry.ww_name.isEmpty())
+            continue; // "Normal" is already added
+        if (entry.ww_id >= ww::stiMax)
+            continue; // Not a format with special name
+        assert(entry.format);
+
+        entry.ww_name = OUString::createFromAscii(ww::GetEnglishNameFromSti(static_cast<ww::sti>(entry.ww_id)));
+        makeUniqueName(entry.ww_name);
+    }
+
+    // 2. Now handle other styles
+    for (auto& entry : m_aStyles)
+    {
+        if (!entry.ww_name.isEmpty())
+            continue;
+        if (entry.format)
+            entry.ww_name = entry.format->GetName();
+        else if (entry.num_rule)
+            entry.ww_name = entry.num_rule->GetName();
+        else
+            continue;
+        makeUniqueName(entry.ww_name);
+    }
+}
+
+OString MSWordStyles::CreateStyleId(const OUString& rName)
 {
     OStringBuffer aStyleIdBuf(rName.getLength());
     for (int i = 0; i < rName.getLength(); ++i)
@@ -340,23 +394,9 @@ void MSWordStyles::BuildStyleIds()
 {
     std::unordered_set<OString> aUsed;
 
-    assert(!m_aStyles.empty());
-    m_aStyles[0].style_id = "Normal";
-    aUsed.insert("normal");
-
     for (auto& entry : m_aStyles)
     {
-        if (!entry.style_id.isEmpty())
-            continue; // "Normal" is already added
-
-        assert(entry.style_id.isEmpty());
-        OUString name;
-        if (entry.format)
-            name = entry.format->GetName();
-        else if (entry.num_rule)
-            name = entry.num_rule->GetName();
-
-        OString aStyleId = CreateStyleId(name);
+        OString aStyleId = CreateStyleId(entry.ww_name);
 
         if (aStyleId.isEmpty())
             aStyleId = "Style";
@@ -589,68 +629,45 @@ void WW8AttributeOutput::DefaultStyle()
     m_rWW8Export.pTableStrm->WriteUInt16(0);   // empty Style
 }
 
-void MSWordStyles::OutputStyle(const SwNumRule* pNumRule, sal_uInt16 nSlot)
+void MSWordStyles::OutputStyle(sal_uInt16 nSlot)
 {
-    m_rExport.AttrOutput().StartStyle( pNumRule->GetName(), STYLE_TYPE_LIST,
+    const auto& entry = m_aStyles[nSlot];
+
+    if (entry.num_rule)
+    {
+        m_rExport.AttrOutput().StartStyle( entry.ww_name, STYLE_TYPE_LIST,
             /*nBase =*/ 0, /*nWwNext =*/ 0, /*nWwLink =*/ 0, /*nWWId =*/ 0, nSlot,
             /*bAutoUpdateFormat =*/ false );
 
-    m_rExport.AttrOutput().EndStyle();
-}
-
-// OutputStyle applies for TextFormatColls and CharFormats
-void MSWordStyles::OutputStyle( const SwFormat* pFormat, sal_uInt16 nSlot)
-{
-    if ( !pFormat )
+        m_rExport.AttrOutput().EndStyle();
+    }
+    else if (!entry.format)
+    {
         m_rExport.AttrOutput().DefaultStyle();
+    }
     else
     {
         bool bFormatColl;
         sal_uInt16 nBase, nWwNext;
         sal_uInt16 nWwLink = 0x0FFF;
 
-        GetStyleData(pFormat, bFormatColl, nBase, nWwNext, nWwLink);
+        GetStyleData(entry.format, bFormatColl, nBase, nWwNext, nWwLink);
 
-        OUString aName = pFormat->GetName();
-        // We want to map LO's default style to Word's "Normal" style.
-        // Word looks for this specific style name when reading docx files.
-        // (It must be the English word regardless of language settings)
-        if (nSlot == 0)
-        {
-            assert( pFormat->GetPoolFormatId() == RES_POOLCOLL_STANDARD );
-            aName = "Normal";
-        }
-        else if (aName.equalsIgnoreAsciiCase("Normal"))
-        {
-            // If LO has a style named "Normal"(!) rename it to something unique
-            const OUString aBaseName = "LO-" + aName;
-            aName = aBaseName;
-            // Check if we still have a clash, in which case we add a suffix
-            for ( int nSuffix = 0; ; ++nSuffix ) {
-                if (std::none_of(m_aStyles.begin() + 1, m_aStyles.end(),
-                                 [&aName](const auto& entry) {
-                                     return entry.format
-                                            && entry.format->GetName().equalsIgnoreAsciiCase(aName);
-                                 }))
-                    break; // Found a unique name
-                aName = aBaseName + OUString::number(nSuffix);
-            }
-        }
-        else if (!bFormatColl && m_rExport.GetExportFormat() == MSWordExportBase::DOCX &&
-                        m_rExport.m_pStyles->GetStyleId(nSlot).startsWith("ListLabel"))
+        if (!bFormatColl && m_rExport.GetExportFormat() == MSWordExportBase::DOCX &&
+                        entry.style_id.startsWith("ListLabel"))
         {
             // tdf#92335 don't export redundant DOCX import style "ListLabel"
             return;
         }
 
-        m_rExport.AttrOutput().StartStyle( aName, (bFormatColl ? STYLE_TYPE_PARA : STYLE_TYPE_CHAR),
-                nBase, nWwNext, nWwLink, GetWWId( *pFormat ), nSlot,
-                pFormat->IsAutoUpdateFormat() );
+        m_rExport.AttrOutput().StartStyle(entry.ww_name, (bFormatColl ? STYLE_TYPE_PARA : STYLE_TYPE_CHAR),
+                nBase, nWwNext, nWwLink, m_aStyles[nSlot].ww_id, nSlot,
+                entry.format->IsAutoUpdateFormat() );
 
         if ( bFormatColl )
-            WriteProperties( pFormat, true, nSlot, nBase==0xfff );           // UPX.papx
+            WriteProperties( entry.format, true, nSlot, nBase==0xfff );           // UPX.papx
 
-        WriteProperties( pFormat, false, nSlot, bFormatColl && nBase==0xfff );  // UPX.chpx
+        WriteProperties( entry.format, false, nSlot, bFormatColl && nBase==0xfff );  // UPX.chpx
 
         m_rExport.AttrOutput().EndStyle();
     }
@@ -699,12 +716,7 @@ void MSWordStyles::OutputStylesTable()
     // Implementing check for all exports DOCX, DOC, RTF
     assert(m_aStyles.size() <= MSWORD_MAX_STYLES_LIMIT);
     for (size_t slot = 0; slot < m_aStyles.size(); ++slot)
-    {
-        if (m_aStyles[slot].num_rule)
-            OutputStyle(m_aStyles[slot].num_rule, slot);
-        else
-            OutputStyle(m_aStyles[slot].format, slot);
-    }
+        OutputStyle(slot);
 
     m_rExport.AttrOutput().EndStyles(m_aStyles.size());
 
diff --git a/sw/source/filter/ww8/wrtww8.hxx b/sw/source/filter/ww8/wrtww8.hxx
index 66866fc9085b..2c79ae8e70c8 100644
--- a/sw/source/filter/ww8/wrtww8.hxx
+++ b/sw/source/filter/ww8/wrtww8.hxx
@@ -1577,10 +1577,12 @@ class MSWordStyles
         const SwFormat* format = nullptr;
         const SwNumRule* num_rule = nullptr;
         /// We need to build style id's for DOCX export; ideally we should roundtrip that, but this is good enough.
+        sal_uInt16 ww_id = ww::stiUser;
+        OUString ww_name;
         OString style_id;
 
         MapEntry() = default;
-        MapEntry(const SwFormat* f) : format(f) {}
+        MapEntry(const SwFormat* f) : format(f) { if (f) ww_id = GetWWId(*f); }
         MapEntry(const SwNumRule* r) : num_rule(r) {}
     };
     std::vector<MapEntry> m_aStyles; ///< Slot <-> Character/paragraph/list style array.
@@ -1589,6 +1591,9 @@ class MSWordStyles
     /// Create the style table, called from the constructor.
     void BuildStylesTable();
 
+    /// Generate proper Word names, taking mapping between special types into account
+    void BuildWwNames();
+
     /// Based on style names, fill in unique, MS-like names.
     void BuildStyleIds();
 
@@ -1603,8 +1608,7 @@ class MSWordStyles
     void SetStyleDefaults( const SwFormat& rFormat, bool bPap );
 
     /// Outputs one style - called (in a loop) from OutputStylesTable().
-    void OutputStyle( const SwFormat* pFormat, sal_uInt16 nSlot );
-    void OutputStyle( const SwNumRule* pNumRule, sal_uInt16 nSlot);
+    void OutputStyle( sal_uInt16 nSlot );
 
     MSWordStyles( const MSWordStyles& ) = delete;
     MSWordStyles& operator=( const MSWordStyles& ) = delete;