bnc#834035 DOCX export: fix hyperlinks of illustration index

We used to export raw Writer bookmarks, but that's not valid in OOXML.
Instead, it has normal bookmarks around the sequence fields, so use them
if they are available.

Conflicts:
	sw/qa/extras/ooxmlexport/ooxmlexport.cxx
	sw/source/filter/ww8/docxattributeoutput.cxx

(cherry picked from commits 750f0ebf97d19d1cf305dabe72d52ad6e90adf70,
e9275c08acc2f4f1c925f78b56a1089515cd9a37 and
8f146a8042dd8f3c88f39f8317372cd9c02ce4ad)

Change-Id: I0ef2ff7967c2802b53752c9505ef6db4cc2b8265
Reviewed-on: https://gerrit.libreoffice.org/5635
Reviewed-by: Jan Holesovsky <kendy@suse.cz>
Tested-by: Jan Holesovsky <kendy@suse.cz>

5 files changed, 103 insertions, 0 deletions

diff --git a/sw/CppunitTest_sw_ooxmlexport.mk b/sw/CppunitTest_sw_ooxmlexport.mk
index e96961fade5b..2e90beca897c 100644
--- a/sw/CppunitTest_sw_ooxmlexport.mk
+++ b/sw/CppunitTest_sw_ooxmlexport.mk
@@ -76,6 +76,7 @@ $(eval $(call gb_CppunitTest_use_components,sw_ooxmlexport,\
     ucb/source/ucp/file/ucpfile1 \
     unotools/util/utl \
     unoxml/source/service/unoxml \
+	uui/util/uui \
     writerfilter/util/writerfilter \
     xmloff/util/xo \
 ))
diff --git a/sw/qa/extras/ooxmlexport/data/bnc834035.odt b/sw/qa/extras/ooxmlexport/data/bnc834035.odt
new file mode 100644
index 000000000000..393c960dea63
--- /dev/null
+++ b/sw/qa/extras/ooxmlexport/data/bnc834035.odt
diff --git a/sw/qa/extras/ooxmlexport/ooxmlexport.cxx b/sw/qa/extras/ooxmlexport/ooxmlexport.cxx
index 2ed291fbe183..fd1c30aea46b 100644
--- a/sw/qa/extras/ooxmlexport/ooxmlexport.cxx
+++ b/sw/qa/extras/ooxmlexport/ooxmlexport.cxx
@@ -10,6 +10,7 @@
 #include <com/sun/star/frame/XStorable.hpp>
 #include <com/sun/star/drawing/FillStyle.hpp>
 #include <com/sun/star/awt/Gradient.hpp>
+#include <com/sun/star/packages/zip/ZipFileAccess.hpp>
 #include <com/sun/star/style/TabStop.hpp>
 #include <com/sun/star/view/XViewSettingsSupplier.hpp>
 #include <com/sun/star/text/XTextFrame.hpp>
@@ -73,6 +74,7 @@ public:
     void testN822175();
     void testFdo58577();
     void testFdo60990();
+    void testBnc834035();
 
     CPPUNIT_TEST_SUITE(Test);
 #if !defined(MACOSX) && !defined(WNT)
@@ -82,6 +84,14 @@ public:
 
 private:
     void run();
+    /**
+     * Given that some problem doesn't affect the result in the importer, we
+     * test the resulting file directly, by opening the zip file, parsing an
+     * xml stream, and asserting an XPath expression. This method returns the
+     * xml stream, so that you can do the asserting.
+     */
+    xmlDocPtr parseExport();
+    void assertXPath(xmlDocPtr pXmlDoc, OString aXPath, OString aAttribute = OString(), OUString aExpectedValue = OUString());
 };
 
 void Test::run()
@@ -123,6 +133,7 @@ void Test::run()
         {"n822175.odt", &Test::testN822175},
         {"fdo58577.odt", &Test::testFdo58577},
         {"fdo60990.odt", &Test::testFdo60990},
+        {"bnc834035.odt", &Test::testBnc834035},
     };
     // Don't test the first import of these, for some reason those tests fail
     const char* aBlacklist[] = {
@@ -144,6 +155,46 @@ void Test::run()
     }
 }
 
+xmlDocPtr Test::parseExport()
+{
+    // Create the zip file.
+    utl::TempFile aTempFile;
+    save("Office Open XML Text", aTempFile);
+
+    // Read the XML stream we're interested in.
+    uno::Reference<packages::zip::XZipFileAccess2> xNameAccess = packages::zip::ZipFileAccess::createWithURL(comphelper::getComponentContext(m_xSFactory), aTempFile.GetURL());
+    uno::Reference<io::XInputStream> xInputStream(xNameAccess->getByName("word/document.xml"), uno::UNO_QUERY);
+    boost::shared_ptr<SvStream> pStream(utl::UcbStreamHelper::CreateStream(xInputStream, sal_True));
+    pStream->Seek(STREAM_SEEK_TO_END);
+    sal_Size nSize = pStream->Tell();
+    pStream->Seek(0);
+    OStringBuffer aDocument(nSize);
+    char ch;
+    for (sal_Size i = 0; i < nSize; ++i)
+    {
+        *pStream >> ch;
+        aDocument.append(ch);
+    }
+
+    // Parse the XML.
+    return xmlParseMemory((const char*)aDocument.getStr(), aDocument.getLength());
+}
+
+void Test::assertXPath(xmlDocPtr pXmlDoc, OString aXPath, OString aAttribute, OUString aExpectedValue)
+{
+    xmlXPathContextPtr pXmlXpathCtx = xmlXPathNewContext(pXmlDoc);
+    xmlXPathRegisterNs(pXmlXpathCtx, BAD_CAST("w"), BAD_CAST("http://schemas.openxmlformats.org/wordprocessingml/2006/main"));
+    xmlXPathRegisterNs(pXmlXpathCtx, BAD_CAST("v"), BAD_CAST("urn:schemas-microsoft-com:vml"));
+    xmlXPathObjectPtr pXmlXpathObj = xmlXPathEvalExpression(BAD_CAST(aXPath.getStr()), pXmlXpathCtx);
+    xmlNodeSetPtr pXmlNodes = pXmlXpathObj->nodesetval;
+    CPPUNIT_ASSERT_EQUAL(1, xmlXPathNodeSetGetLength(pXmlNodes));
+    if (aAttribute.isEmpty())
+        return;
+    xmlNodePtr pXmlNode = pXmlNodes->nodeTab[0];
+    OUString aValue = OUString::createFromAscii((const char*)xmlGetProp(pXmlNode, BAD_CAST(aAttribute.getStr())));
+    CPPUNIT_ASSERT_EQUAL(aExpectedValue, aValue);
+}
+
 void Test::testZoom()
 {
     uno::Reference<frame::XModel> xModel(mxComponent, uno::UNO_QUERY);
@@ -709,6 +760,18 @@ void Test::testFdo60990()
     CPPUNIT_ASSERT_EQUAL(sal_Int32(0x00FF00), getProperty<sal_Int32>(getRun(xParagraph, 1), "CharColor"));
 }
 
+void Test::testBnc834035()
+{
+    // This is tricky, when saving manually, there are 2 hyperlinks, here only
+    // one, no idea why. That one still shows that we're not using bookmarks, though.
+
+    // Illustration index had wrong hyperlinks: anchor was using Writer's
+    // <seqname>!<index>|sequence syntax, not a bookmark name.
+    xmlDocPtr pXmlDoc = parseExport();
+    // This was Figure!1|sequence.
+    assertXPath(pXmlDoc, "/w:document/w:body/w:p/w:hyperlink", "anchor", "_Toc363553908");
+}
+
 CPPUNIT_TEST_SUITE_REGISTRATION(Test);
 
 CPPUNIT_PLUGIN_IMPLEMENT();
diff --git a/sw/source/filter/ww8/docxattributeoutput.cxx b/sw/source/filter/ww8/docxattributeoutput.cxx
index 08d79f02118b..0e2ee5e8969e 100644
--- a/sw/source/filter/ww8/docxattributeoutput.cxx
+++ b/sw/source/filter/ww8/docxattributeoutput.cxx
@@ -699,6 +699,7 @@ void DocxAttributeOutput::DoWriteBookmarks()
             FSNS( XML_w, XML_id ), OString::valueOf( sal_Int32( nId ) ).getStr(  ),
             FSNS( XML_w, XML_name ), rName.getStr(),
             FSEND );
+        m_sLastOpenedMark = rName;
     }
     m_rMarksStart.clear();
 
@@ -828,6 +829,12 @@ void DocxAttributeOutput::StartField_Impl( FieldInfos& rInfos, bool bWriteRun )
 
 void DocxAttributeOutput::DoWriteCmd( String& rCmd )
 {
+    OUString sCmd = OUString(rCmd).trim();
+    if (sCmd.startsWith("SEQ"))
+    {
+        OUString sSeqName = msfilter::util::findQuotedText(sCmd, "SEQ ", '\\').trim();
+        m_aSeqMarksNames[sSeqName].push_back(m_sLastOpenedMark);
+    }
     // Write the Field command
     m_pSerializer->startElementNS( XML_w, XML_instrText, FSEND );
     m_pSerializer->writeEscaped( OUString( rCmd ) );
@@ -1313,8 +1320,33 @@ bool DocxAttributeOutput::StartURL( const String& rUrl, const String& rTarget )
             m_pHyperlinkAttrList->add( FSNS( XML_r, XML_id), sId.getStr());
         }
         else
+        {
+            // Is this a link to a sequence? Then try to replace that with a
+            // normal bookmark, as Word won't understand our special
+            // <seqname>!<index>|sequence syntax.
+            OUString aMark(sMark);
+            if (aMark.endsWith("|sequence"))
+            {
+                sal_Int32 nPos = aMark.indexOf('!');
+                if (nPos != -1)
+                {
+                    // Extract <seqname>, the field instruction text has the name quoted.
+                    OUString aSequenceName = OUString('"') + aMark.copy(0, nPos) + OUString('"');
+                    // Extract <index>.
+                    sal_uInt32 nIndex = aMark.copy(nPos + 1, aMark.getLength() - nPos - sizeof("|sequence")).toInt32();
+                    std::map<OUString, std::vector<OString> >::iterator it = m_aSeqMarksNames.find(aSequenceName);
+                    if (it != m_aSeqMarksNames.end())
+                    {
+                        std::vector<OString>& rNames = it->second;
+                        if (rNames.size() > nIndex)
+                            // We know the bookmark name for this sequence and this index, do the replacement.
+                            sMark = OStringToOUString(rNames[nIndex], RTL_TEXTENCODING_UTF8);
+                    }
+                }
+            }
             m_pHyperlinkAttrList->add( FSNS( XML_w, XML_anchor ),
                     OUStringToOString( OUString( sMark ), RTL_TEXTENCODING_UTF8 ).getStr( ) );
+        }
 
         OUString sTarget( rTarget );
         if ( !sTarget.isEmpty() )
diff --git a/sw/source/filter/ww8/docxattributeoutput.hxx b/sw/source/filter/ww8/docxattributeoutput.hxx
index 0854e46ea662..85014b9ca97f 100644
--- a/sw/source/filter/ww8/docxattributeoutput.hxx
+++ b/sw/source/filter/ww8/docxattributeoutput.hxx
@@ -603,6 +603,13 @@ private:
     /// Maps of the bookmarks ids
     std::map<OString, sal_uInt16> m_rOpenedMarksIds;
 
+    /// Name of the last opened bookmark.
+    OString m_sLastOpenedMark;
+
+    /// If there are bookmarks around sequence fields, this map contains the
+    /// names of these bookmarks for each sequence.
+    std::map<OUString, std::vector<OString> > m_aSeqMarksNames;
+
     /// The current table helper
     SwWriteTable *m_pTableWrt;